YT YouTube2Text

YouTube2Text → Transcripts → [CS61C FA20] Weekly Lecture 09.LIVE - Cache

[CS61C FA20] Weekly Lecture 09.LIVE - Cache — Transcript

by CS 61C Departmental · 17,397 words · 2,575 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Sure, not to.
0:02Forget that and I'm going to then also
0:04grab I forgot to change my name, but I'm
0:05going to do this. I'm going to spotlight
0:08New Bora and I'm going to spotlight
0:09myself and then the rest of the folks
0:11won't be spotlit on that and add that
0:13spotlight. There's just two of us and
0:15then we'll spotlight our speakers our
0:16guests when we when we share that.
0:18Welcome students. Welcome everybody. I
0:19see the numbers are
0:21rolling in here. Perfect.
0:23And let me forget how to change my name.
0:27Wonderful.
0:29Yeah, don't go and incriminate her.
0:34All right.
0:45All right, I think we're ready to start.
0:48Ladies and gentlemen, welcome.
0:50Welcome everybody to I'm getting a Let
0:53me
0:53I'm getting a little bit of a there.
0:55Okay. Welcome everybody to week I think
0:57we're week nine CS 61C live lecture.
1:01Woo!
1:02Good to see everybody. Intro to caches.
1:05We're doing caches this week a very
1:06exciting time. Um lots of things are
1:09happening 61C. We want to make sure to
1:10share that and just jump right in. We
1:11don't want to delay. So, let me go to
1:14the next slide.
1:16Here's the agenda what we're going to
1:17do. We're going to we're past the
1:19halfway point. We're downhill from here.
1:21So, that's just wonderful. Um it's it's
1:23nice to feel like Thanksgiving is not
1:26that many weeks away.
1:27Looking forward to that. We're going to
1:28talk about computing in the news.
1:29There's some exciting I'm going to show
1:30a video for that. We have two wonderful
1:32guests from Intel here. We'll update our
1:34schedule as always. We'll do a sanity
1:36check to actually share with you how our
1:37students are doing. We have some good
1:39news for you all. So, stay for that. We
1:41hope that you're all uh share this
1:42information with all the rest of the
1:43other uh 1,100 people who aren't part of
1:46this class, but we're recording this so
1:47we'll share with that. We have a very
1:48important relief package that we want to
1:51share. We were able to get this through
1:52Congress. So, we have the official 61C
1:54relief package we want to talk about and
1:56make sure you stay for that. 10-point
1:58relief package. We want to talk about
1:59the quest retake and update on the
2:02midterm which is coming up this coming
2:03weekend and we'll have an ask us
2:04anything if we have time at the end.
2:05Okay? And that's always what we try to
2:07do.
2:08All right. So, I have a video I want to
2:10show you. This is on neuromorphic
2:12computing. This is a one of the Intel's
2:13new videos and I want to see if I can
2:14share this with sound if I get this done
2:16right. Here we go.
2:18So, Google Chrome with sound and give me
2:21the thumbs up Bora if this is working
2:22for you. Ready?
2:25Organic.
2:26Intuitive.
2:28Fluid.
2:29Neuromorphic computing is designed to
2:31work less like today's microprocessors
2:34and more like the human mind. Because
2:37brains are so different than transistor
2:39circuits, we have to rethink everything
2:41we know about computer architecture.
2:44Traditional microprocessors are
2:46synchronous, meaning every function is
2:49governed by a clock. The clock says when
2:51instructions are executed and when
2:53functions happen. Everything works in
2:55unison according to the clock.
2:58Our brains are different. They're
3:00asynchronous, meaning no internal clock.
3:03Neurons operate with no prescribed
3:05order.
3:06We just have thoughts.
3:08Thoughts that build on each other
3:10or split off into different ideas.
3:13That's the way neuromorphic computing
3:15works, too. It's completely
3:17asynchronous.
3:18So, functions and instructions happen
3:21when they happen in parallel.
3:23Which means neuromorphic systems think
3:25more like brains do.
3:27They can adapt, recognize nuances, and
3:30teach themselves as they work. It's an
3:32elegant,
3:34organic, and beautiful way to think.
3:41All right.
3:43Be sure to like and subscribe. So, let
3:45me
3:47Let me get out of that now. All right.
3:49And uh get back to this.
3:52And let's you talk about that. Bora,
3:54take us take us to the some of that
3:56work.
3:56Yeah.
3:57So, one thing that we wanted to that you
4:00wanted to wanted to talk about
4:02today is in the past few lectures we
4:05have talked about traditional
4:06microprocessor that is operating
4:09under the principles of a synchronous
4:11data system.
4:13But there is a lot of stuff that is
4:15happening now that is labeled as
4:17neuromorphic computing and it is a
4:20little bit different in what how we do
4:21things. Uh neuromorphic doesn't
4:23necessarily mean that everything is
4:25asynchronous, but people have been
4:27building asynchronous chips such that
4:29they look uh a little bit more like
4:32human brains.
4:34Uh processor cores that are in these
4:37chips are
4:38perhaps of the complexity of a
4:4161C project. So, you know, if you jump
4:43into this you might be able to build one
4:44of these.
4:45So, Intel published a couple of years
4:48ago um this 128 core neuromorphic chip
4:52in 14 nanometer CMOS. Each of these
4:54cores is not much more complicated than
4:58um
4:59you know, in in
5:00when we look at the logic not much more
5:02complicated than 61C project. I mean, at
5:04the first glance there is a lot of stuff
5:06that happened behind it. It has, you
5:08know, more memory and things like those.
5:10But then there's been a lot of stuff now
5:11then when they're trying to put a
5:13neuromorphic computing system
5:15out of out of that.
5:17Um so, they have all these systems. The
5:20boards have Hawaiian names. So,
5:23um you know, there is Loihi,
5:25uh Hokulani, Pohoiki. I I somebody is
5:28inspired by Hawaiian names uh here, but
5:31you know, in the end this is a pretty
5:32big machine
5:33uh that Intel is building.
5:35So, we have a great privilege today that
5:38to have two guests from Intel.
5:40Uh both of them um
5:43probably didn't have exactly the most
5:45linear paths into that careers and you
5:48know, we do recognize that
5:50you know, two thirds of this class are
5:51not necessarily ECS majors. Many of them
5:54are Many of you are not even um L&S
5:58uh CS majors.
5:59Coming with all kinds of different
6:00backgrounds. So, I I'm going to ask uh
6:04um
6:05uh Farhana, Dr. Farhana Sheikh, and uh
6:08um
6:09uh Dr.
6:11Srinivas to introduce themselves to tell
6:13us, you know, how they ended up doing
6:15what they do now. Uh Farhana, I don't
6:17think he's working on neuromorphic now,
6:19but she can tell us how she ended up
6:21um
6:22at in Intel research.
6:24Um and then we can chat a little bit
6:25about neuromorphic and other interests.
6:27So, Farhana, tell us.
6:30Okay. So, I did my bachelor's in system
6:33and computer engineering long time ago.
6:36So, this was a kind of a broad program
6:37with a very with a small emphasis on
6:39circuits, uh but not too much. It was
6:42mostly at the system level and the
6:43software level
6:45and and architecture level. And then I
6:48ended up um coming to Berkeley uh for a
6:51little bit, did my master's in in with
6:55Edward Lee and we kind of talked about
6:57more on the software side of things and
7:00and I worked on Ptolemy and then I went
7:03back to work. I worked in a CAD company
7:06and I worked there for 5 years. I became
7:08a senior manager. I was actually
7:10managing managers and then I realized
7:12that I wanted to come back to school.
7:13So, I came back to Berkeley again and I
7:16started off in the CAD field and then I
7:18joined Bora's team and I took some
7:21undergrad classes in circuits and I also
7:24took some CS classes as well in the
7:26undergrad and
7:28and then I actually changed my field to
7:29circuits. So, that's how I ended up in
7:31circuits and then I went to Intel.
7:34Yeah, one day I may be known as
7:37Farhana's former advisor.
7:39That was my
7:41my biggest accomplishment. Um
7:44Naren, you also had a
7:46probably not the most direct path
7:48to to being an architect of Loihi.
7:52You're muted. You're muted.
7:57My path is actually even more strange
7:59than
8:00than you normally imagine. So, I started
8:02as a mechanical engineer
8:04um my undergrad in India and then
8:07subsequently my PhD from here
8:10and
8:11um mostly focused on robotics and
8:13adaptive control as the areas.
8:16Um
8:17Uh but but the
8:19uh
8:19as you know,
8:21uh these areas
8:22have emerged tremendously to the point
8:25where uh you know, of course that the
8:26heart of all this is the computer and uh
8:29so, being a curious person I've always
8:31wanted to learn
8:32just more than just the the robot and
8:34the associated task, but what what
8:36drives them? What is the underlying uh
8:39hardware that
8:40uh is implemented that allows us to
8:42actually do what we do.
8:44And so, I started learning a lot more
8:46about it through courses first during my
8:48PhD
8:49uh endeavor and then uh
8:51I actually had some hands-on experience
8:53at Urbana-Champaign when I did a
8:55post-doctoral fellowship there
8:57working on visual processing um
9:00and uh you know, signal processing uh
9:03kind of areas.
9:04Uh used to call it human computer
9:06intelligent interaction those days.
9:08Uh there was a big army funded research
9:10lab there
9:12where we started working on some of
9:13those problems
9:14and it became apparent that um
9:17more and more the issues are more
9:19challenging in from the perspective of
9:20energy efficiency. That is how do you
9:22make this hardware um
9:25sort of do what uh humans do from the
9:27perspective of processing information
9:30and
9:31that is the efficiency with which they
9:32process information. And so,
9:34that
9:36uh stint in uh UIUC then uh let me to
9:39Hughes Research Labs which is in Malibu,
9:41California. You may may or may may not
9:43know about it, but it's a corporate lab
9:45funded by Boeing and uh General Motors
9:48as owners, but they do a lot of um
9:50research for the government um
9:53and there was a lot more challenging
9:55problems when I went there because now
9:58the government as you know is always
10:00interested in the leading edge uh in
10:02terms of efficiency because many of the
10:04assets as they call it are deployed
10:06where the the power is a big question.
10:09Energy efficiency is a huge problem. And
10:11so uh that led my experience to sort of
10:15drift towards designing architectures
10:18that would be more efficient. And
10:21uh guess what? I mean brain our brains
10:23are a great example of
10:26of trying to sort of understand what's
10:27going on there. And so I got interested
10:30in that area. And
10:32uh have ever since sort of worked
10:35heavily on that. I started my first
10:37opportunity came 2004 when I worked on a
10:40DARPA-funded program called biologically
10:43inspired cognitive architectures.
10:46So that allowed us to sort of figure out
10:48various um computer elements that go
10:51into an architecture for doing efficient
10:53processing. And from then on had the
10:56fortune of working on several other
10:57DARPA programs like HRL including one
11:01called Synapse, which is a very famous
11:03uh neuromorphic computing program that
11:05was uh that was uh basically the outcome
11:08of which was a true north chip from IBM
11:11uh and uh as part of HRL we did a small
11:14chip that did some learning as well.
11:16And then uh Intel came calling and I
11:19they said, "Well, you know, we like what
11:20you're doing uh the HRL chip. Uh would
11:22you come uh to Intel and help us design
11:25uh uh the first generation of our Intel
11:28uh neuromorphic hardware?" And that was
11:30quite exciting for me. And as you can
11:32see my transition from a mechanical
11:34engineer to a hardware architect and uh
11:37sort of looking at the spectrum of
11:39problems along the way helped me sort of
11:41transition from one to the other. And uh
11:43today I am more known for my work in the
11:46neuromorphic area than uh any of my past
11:49work.
11:50So that's how uh my career shaped up to
11:53become more aligned with the uh you
11:56know, computer architectures if you
11:57will. Yeah.
11:59Thank you. That's awesome.
12:02Oh, I have an echo so much. Um can you
12:05tell us a little bit about what is
12:06neuromorphic computing a little in you
12:08know, just a little bit more of a
12:10resolution than what I said?
12:16You're muted. You're muted. We you have
12:19an echo when you're not speaking so I
12:20had to mute you. Sorry.
12:21Yeah.
12:22Oh, sorry.
12:23Um okay. So um basically a neuromorphic
12:26computer uh you know, you can think of
12:28it as uh like you like like the video
12:30says, you know, there's no clock in the
12:32system.
12:33Furthermore, the instruction set is
12:35mostly based on you know, what a neuron
12:38and a synapse do in terms of compute. So
12:40basically you have this multi-core
12:42architecture where each core houses a
12:45bunch of neurons and synapses that are
12:47basic computer elements and the
12:49operations of which are are you know,
12:51sort of dictated by by um
12:54instruction sets that are not exactly
12:58does not readily map to what you see in
13:00a typical computer.
13:02Um and then like you pointed out there's
13:05a this is a multi-core system so there's
13:07an asynchronous computer element here
13:09where each neuron integrates information
13:12and sends out its own sort of a pulse
13:14that dictates how strong the information
13:17or the input was to that particular
13:18neuron. And then there's a stream of
13:20pulses and they're all acting in
13:22parallel. So it's about handshaking
13:24between neurons and when they receive
13:25information to process what they get and
13:28then pooling that information across
13:30other neurons to sort of make inferences
13:32about what's going on. Uh so thereby
13:35sort of avoiding this need for a
13:38global clock
13:39uh that you see typically see in a
13:42in a computer architecture. Uh
13:44synchronous architecture today. Uh so
13:47uh notwithstanding that how do you
13:50design algorithms and and software that
13:52allows you to run applications that are
13:54useful given this paradigm is on the
13:56other side of the story that we worked
13:58on and um
14:00what it turns out it's sort of like a
14:02you know, this the class of algorithms
14:04that it actually works well for is sort
14:06of complimentary we find to the ones
14:09that we normally are used to from the
14:11from the Intel side of things. The
14:13general purpose computing versus this
14:15one it's you know, things like graph
14:17search or things like um
14:20sparse coding uh you know, uh things
14:22like constraint satisfaction. These are
14:24all problems uh domains that are sort of
14:27um
14:28up up and coming but they're sort of
14:29orthogonal to your normal machine
14:32learning MAC operations kind of things
14:33that you see today. And they complement
14:36each other pretty well and so Intel is
14:38very interested in figuring out that
14:39space of uh compute
14:42domain that neuromorphic can offer. Uh
14:45it's uh energy efficiency for and then
14:48uh see how we can go build uh chips that
14:52people would be interested in using. So
14:55that's where things are now. Yeah.
14:57Can I ask a just real brief question? Or
14:59I didn't I didn't mean to to short
15:01circuit you you were doing the interview
15:02today but um
15:03the students just finished a a project
15:06we called the game of life. You know,
15:07you have a a grid of black and white
15:09cells and they live or die and then you
15:11have the next like it's a very much a
15:13heartbeat kind of problem. Is this and
15:15this is it's an array and then there's
15:16the next day and the next day. How would
15:19that kind of a problem be encoded in
15:20your system? How do you I mean you can't
15:22just have them all random because they
15:24depend on each other and it's like you
15:25have to take this day minus one affects
15:28day I, you know, and then the day the
15:29next affects the day n plus one. So how
15:31do you do that in a asynchronous way?
15:34Can you help us with that?
15:35Yeah, that's a good question. So um so
15:38you know, think of it like a Sudoku is
15:40one example like constraint satisfaction
15:42problem. Sudoku would be one where you
15:43have a state and then you do to make a
15:46move and then it transitions, right? So
15:49uh the way this so the way we we are our
15:52system is architected is you don't have
15:55to do
15:56um
15:57so each neuron let's let's say
15:59represents one particular node in the
16:01Sudoku board or in a game of life
16:04uh thing. And they sort of locally
16:06integrate information from everybody
16:08else that they communicate with. For
16:09example, that the 3 by 3 square in a
16:12Sudoku, a neuron uh would sort of look
16:15at the states locally there and pool and
16:17to sort of infer what it needs to do
16:19based on the constraints it needs to
16:20satisfy.
16:22Um so that can happen sort of
16:24asynchronously to the another square
16:26next to it or the one after that as long
16:29as the you know,
16:30there's a communication that allows you
16:32to do a global constraint at some point,
16:34some checks at some point. So um you can
16:37sort of see a sort of a demarcation of
16:40asynchrony versus synchrony uh even in
16:43that problem. You don't have to all
16:45compute it with a clock but you can sort
16:46of break it down. And especially if it's
16:49more parallel then you can sort of see a
16:51combination of the two in that case.
16:53Um
16:54So
16:55yeah, I don't know if that answered it
16:56or not but yeah. No, that makes sense. I
16:57can almost imagine a game of life where
16:59you have two different areas with you
17:00know, where there's like ones you know,
17:02you know, the alive creatures are
17:03different areas and this because
17:05they don't talk to each other and
17:06there's a gap and so this might actually
17:08be going faster and be able to like this
17:10might be a generation a thousand, this
17:12is still in generation five because it's
17:13more stuff to compute. So it's a it's an
17:15asynchronous model. And if they ever
17:17were to bleed together, the guys in the
17:18middle would say, "Look, I'm not ready
17:20to have my final state until you're both
17:21left and right are at the same day."
17:23That's right. And if I am day a
17:24thousand, I need both guys day a
17:26thousand to compute day a thousand plus
17:27one. And if you're day a hundred, I'll
17:29wait for you to get to get to day a
17:30thousand and now that you're at the same
17:31place, now I can do a thousand plus one.
17:33Exactly. That makes a lot of sense.
17:35Really interesting. Boy, that's really
17:37really interesting.
17:38Perfect.
17:38All right. Let's
17:40um
17:41have I know what Farhana is doing. I
17:43mean she does a lot of stuff with Can
17:45you tell us what you're
17:46working on? So I just know what you're
17:48working on Farhana.
17:53You need to unmute yourself.
17:55Yeah, yeah. Sorry. So um so yeah, so I'm
17:58I'm actually come first full circle. So
18:01uh I'm actually working on um
18:03heterogeneous systems integration. So I
18:05was a system engineer a long time ago.
18:08Um and now what I'm doing is I'm
18:10actually working from the circuits level
18:12and I'm actually a little bit deeper
18:15into the technology side of it and
18:17trying to figure out how to
18:19um advance our technology to support
18:22heterogeneous integration. So 3D
18:24stacking is really something that I'm
18:26looking at
18:27right now uh in in my current job.
18:31And so I'm trying to stack um you know,
18:34uh RFICs so wireless chips with uh
18:37digital architectures and digital
18:39processors and and and FPGAs. So that's
18:42my current work. So it actually requires
18:44you to understand software. It requires
18:47you to understand architecture. It
18:49requires you to understand technology
18:51and circuits.
18:53And can you say like in in a few
18:56sentences what is heterogeneous
18:58integration uh I mean so so um a typical
19:01sophomore uh
19:03can understand? Okay. So heterogeneous
19:06integration is really taking thing that
19:09things that are disparate like things
19:10that are different and and trying to put
19:13them into a single package. For example,
19:15um you would you could have a processor
19:19uh but at the same time you would want
19:20your processor to say um interact with a
19:24wireless chip so say a modem or
19:26something. And you'd want to bring those
19:28two things very close together.
19:30Typically we would have processors
19:31located on a separate package and and a
19:34and a modem on a separate package but
19:36you want to bring them together in the
19:37same package. So that's what we call
19:39heterogeneous integration. And actually
19:41if you open up your Apple iPhone and you
19:44deconstruct it, you'll actually see one
19:46of the biggest examples of heterogeneous
19:48integration in there.
19:51Is that related a little bit to this
19:54thing about Moore's law may not be with
19:57us forever? So, I think some students
19:59maybe worried about the health of of
20:01Moore's law these days. And this is is
20:04this a relief that we do not have to
20:07keep pounding on Moore's law forever?
20:09Uh that Yeah, so you know, a lot of
20:12a lot of people are are concerned that
20:14you know, Moore's law is running out of
20:16steam and we're not going to get to
20:17smaller nodes. So, what's going to
20:18happen?
20:19Uh so, really Moore's law actually
20:21mentions this in in his 1967 paper where
20:25he says that you know, at some point
20:27what we're going to be doing is we're
20:28going to be working on putting uh
20:30different uh chips together, different
20:32chips that do different things. And
20:33we're going to be you know,
20:35putting them together in single
20:36packages. And so, this is kind of
20:39extending Moore's law beyond just
20:41technology scaling.
20:44Perfect. Thank you. So, um
20:46there was one question
20:48uh
20:49but I'm going to ask some some questions
20:51on on behalf of
20:53of students. One question is
20:55where do you work? You know, are you in
20:57headquarters in Santa Clara or somewhere
20:59else? I think I know where Farhana is. I
21:01mean, you can see the hills behind her.
21:07Yeah, I'm also based out of Santa Clara.
21:09Yeah, so I'm I'm I'm I'm in I'm in
21:11Portland, Oregon.
21:13And and the interesting thing is that
21:14Portland, Oregon has the majority of
21:16Intel labs there. But Intel labs is
21:18spread across Santa Clara and and China
21:21and India.
21:23And I was in Intel labs for 11 years. I
21:25recently moved over to the Altera side
21:27of the business, the FPGA side of the
21:29business. And I work in the CTO office
21:31there. But and so, all my team is in San
21:33Jose, but I'm in Portland. And I get to
21:37when you know, post COVID, I'll get to
21:39go back on the plane every week to to
21:41visit my team.
21:43Oh, oh, she said CTO's office. So, you
21:46know, she's not
21:48It's called the it's called the FPGA
21:50Yeah, it's called the FPGA CTO office.
21:53Yeah. Um
21:55A question that we typically ask do you
21:57guys hire interns? You know, and and
21:59what would it take to if you're a
22:01sophomore in Berkeley in
22:03computer science or electrical
22:05engineering computer science to get an
22:06internship at Intel? I mean, is there a
22:09You know, how do we go about that?
22:12Uh Narayan, do you want to go first?
22:13Yeah, sure. Yeah. So, actually Bora, we
22:16have
22:17what is called as an
22:18Intel neuromorphic research community or
22:21INRC.
22:23So, this is basically set up to work
22:25with
22:26with the academic community and students
22:29to sort of help um
22:31uh look into problems this problem space
22:33of neuromorphic computing and see uh
22:36you know,
22:37um
22:38we can encourage students to sort of
22:39learn about it as well as to um apply
22:42their own ideas to create new uh
22:45possible applications uh given the
22:48hardware we develop.
22:49Um so, there are opportunities every
22:52year that we look into for you know,
22:54hiring some students as interns to sort
22:56of
22:57explore some space uh of interest that
22:59Intel may have. Um So, if there's
23:02interest, I can send you a link
23:04uh to sort of share with the students.
23:06So, they can uh you know, definitely uh
23:09reach out to us and we'll be happy to
23:11look into it.
23:13Yep. And and I can
23:14send us send us a send us a
23:17a link and we're going to put it on
23:18Piazza right away. Yeah.
23:20Okay.
23:20Yeah, and I I just want to follow up
23:21with that is that yes, we have uh
23:23opportunities at all levels. We have
23:25opportunities at the undergraduate
23:26level. We are interested in in uh you
23:30know, folks that are doing their
23:31bachelor's and coming on board and
23:33working at Intel. And we also have
23:35opportunities at the graduate level,
23:39Yeah, and and what I think I one thing
23:41that I wanted to add last fall, I was
23:43teaching 151. Um
23:46there was a myth at some point that you
23:48need to have an advanced degree in order
23:49to be hired by you know, these hardware
23:52companies. I can tell you, the largest
23:54number of people was uh were with
23:57bachelor's level were hired by Apple.
23:59The second largest were were went to
24:01Intel. Yep. Um so, it's totally
24:04possible. I mean, you just you know,
24:05it's very doable.
24:07Um but we need to get work on these in
24:09internships uh a bit more. Uh they're a
24:12big stepping stone. So, if you guys have
24:14an internship program that you can share
24:16with us, send us send us over.
24:18Definitely. With that,
24:20I that doesn't look like there are any
24:21other questions. I think we have sorted
24:23out most of the stuff. Let's thank uh uh
24:26Farhana and Narayan.
24:28Uh thank you very much. Thank you.
24:30Uh absolutely. Our pleasure. Thank you
24:31for inviting us. Thank you.
24:32Yeah, thank you very much. It's good to
24:34be here. Thank you for joining us.
24:36Wonderful.
24:37Thank you.
24:39All right. And we are going to move on.
24:42You feel free to All right, schedule.
24:45Here we Here we go.
24:47So, um Bora, you want to cover that
24:49while I'm doing the spotlight thing real
24:51fast? Yeah, jump on in if you want.
24:53Yeah, uh let me let me see. Hang on.
24:56Hang on. Hang on.
24:57Today, I'll I'll just say we're we're at
24:59uh week nine, Monday week nine, 10/19.
25:02It's the last lecture on pipelining.
25:03You're closing off the idea that we're
25:05making a computer faster. No different
25:08in the output, just a little bit more
25:09efficient in terms of being able to do
25:10multiple things at once at the pipeline
25:12stages. Um then my set of lectures comes
25:15in and we're going to have four lectures
25:16on caches. And very important, this
25:19cache information is not on the midterm.
25:22So, you can stop worrying about the
25:24midterm after Monday's material and then
25:26just focus for four days on the midterm.
25:29And again, the midterm is Sunday and
25:30Monday. You have 48 hours to do that.
25:32So, you've got kind of like a week, a
25:33full you know, five or six days to to
25:35work on the stuff that you're you need
25:36to practice on before then.
25:39There's no lab this week.
25:40Yeah, so what I wanted to
25:41Yeah, so what I wanted to add about that
25:43Today's lecture is
25:45it was posted last week but it is only
25:48about 35 minutes. I mean, we I it just
25:50basically ties the loose ends in in in
25:52pipelining. And it is all what you need
25:55to complete both project uh 3B
25:58and uh the last part piece of a puzzle
26:01for the midterm.
26:02Um so, you you and there is no also
26:04there is no lab this week. So, try to
26:06use that time to to fully digest uh this
26:09material.
26:12Sounds good.
26:14Let's see. And the midterm material goes
26:15up to up to Monday here. Um and keep
26:18keep hearing the questions up. We're
26:20We've got a lot I want to make sure to
26:21get to the next piece. This is a really
26:22important piece.
26:24Bora and I did a So, let's see.
26:26You want to talk about it? Go ahead. No,
26:28well, we did a sanity check to see to
26:29see who how you were doing. We just
26:32wanted to say, you know, in the in the
26:33health and wellness you know, workload
26:35wellness, that's one element of it. But
26:37the other is performance. How are you
26:38performing in this class? So, I just
26:41wanted to give a sanity check and then
26:42Bora can give and somehow I have
26:44This is a it's green. So, my green is
26:47showing Berkeley is like a Berkeley. I
26:49love this. This is like
26:50basking at the Berkeley picture.
26:53There's a lot of purple in that picture.
26:55Folks are doing really really well. I
26:57think if I turn off my if I if I adjust
26:59this, I'd adjust it. But folks are doing
27:01really really well on project one
27:03uh in terms of that. Just really really
27:05doing well. Um
27:07and very very pleased with project 1A.
27:10Just really really pleased with 1A. So,
27:12So, I I what I wanted to say, I mean,
27:14what does it mean when the the mean
27:17mean is
27:19when the median is a 100 points out of
27:21100 points, that means a little over
27:23than half the class is getting A pluses.
27:27Um
27:29For this assignment. For this for this
27:30piece. For this particular assignment.
27:32Yeah, yeah. Uh we can take a look at the
27:34next one. Yeah, let's Here we go. Next
27:36one. Let's see how we're doing. How's
27:37the class doing? Here's 1B.
27:40Wait, it didn't change. You're right. It
27:42didn't change. Look, it even got better.
27:45Look,
27:45okay, maybe the mean is a little lower.
27:48But I mean, out of you know, what do we
27:49have total students? 1,100 maybe
27:51students? 970 are acing it. I mean, just
27:54remarkable how proud we are about folks
27:56putting the work in and getting it. So,
27:57you're doing a great job.
27:58Let's move on to the next one. All
28:00right, let's see.
28:02That's 2A.
28:04The median is still 100%.
28:06By the way, I don't think this is usual,
28:08by the way. This might seem usual to
28:09you. This isn't usual. Normally, there's
28:11a distribution.
28:12People are just crushing it. You may
28:14maybe we're adjusting and pulling it
28:15back, but people are crushing it this
28:16year in a way that I haven't seen
28:18before. If I pull up I need I'm going to
28:19do some deep data to look at like last
28:21year and a year before and a year
28:22before. I don't remember seeing these
28:24numbers. I really don't remember seeing
28:25these numbers. So, it's it's remarkable
28:27seeing how much success students are
28:29doing here.
28:30So, so we are very pleased with this. I
28:32mean, we can move on to project 2B where
28:35there was some discomfort about how
28:36things were were graded. We have a new
28:38distribution. The distribution has
28:40changed a lot in the next one.
28:41go. Here's the 2B.
28:44And this changed a lot. This changed a
28:46lot after we took a look at the
28:48autograder and Steven put a lot of
28:50effort into thinking about the points.
28:51And I think we were initially too harsh.
28:53And I thank you for that feedback. I
28:54really want to thank folks for the
28:55feedback about uh 2B.
28:59Uh and we've I think we've corrected it.
29:01Like, this is what you by the way this
29:02kind of graph is what you normally see
29:03for 1A, 1B, and 2A. We normally see this
29:06kind of shapy thing. But it's normally
29:07you know, it's heavily weighted towards
29:09the hot the higher hump. But you don't
29:10see people acing it. So, I love the fact
29:12that people are just really crushing the
29:14first three parts of those things. Just
29:16great job, folks. Yeah, and so, the big
29:18news here is you know, before we before
29:21Steven spent many hours last night in in
29:24trying to re you know, assign partial
29:27credits
29:29to
29:30to
29:31the the parts of a pro project that were
29:34uh failing on on uh
29:37uh um
29:39um
29:41The failing test. Failing the hidden
29:42test.
29:43yeah, yeah, yeah, yeah, hidden test. Um
29:46on on uh conventions. Um,
29:50the
29:52big thing here was the we had 150
29:55something students with new zeros or
29:58close to zeros and that definitely did
30:00not feel good. So, we tried to
30:03assign partial credits to that. That
30:05dramatically that didn't change much the
30:08our
30:09um
30:09um
30:12mean
30:14um hang on here. Yeah, that didn't
30:16dramatically change our mean but it
30:17shifted the median
30:19by a lot. By by you know 20 points or
30:22so. So, this looks more normal. Now, if
30:25we exclude
30:27um
30:28the the quest because there are still
30:30two more opportunities to claw back the
30:32quest. So, you know, this isn't you know
30:34in you know first interim score on quest
30:37that we have now. At the moment, you
30:40know, based on what is happened the
30:42grades that have been assigned so far,
30:45we have about 2/3 of a class that are
30:46getting A's.
30:48Um, and
30:492/3 that's never happened before. That
30:50has never happened before at week week
30:52eight that 2/3 of students are getting
30:54A's. And by the way, you know Bora and
30:56my spirit of this, we want to give
30:57everybody an A. We think if if you earn
30:59an A, you will get an A and we will
31:01celebrate that. There's there's no curve
31:02in this class. This is an absolute
31:03graded class and folks are doing really
31:06really well. And the the question on the
31:07thing was getting A's and what getting
31:08A's and all the if you just look at the
31:10projects, if we ignore the quest for
31:12now, the class so far ignore because the
31:14quest really was you know, not a typical
31:16quest. We think the midterm will do much
31:17better and people will then override the
31:19quest and it'll just be a bad memory.
31:21So, or we'll all go back and laugh
31:22together. We'll you know, share a
31:23sparkling water and we'll
31:26clink our glasses together and and that.
31:27But I think the students are doing
31:29remarkably well right now. Really
31:31remarkably well. Although that said,
31:33there's a chunk of students that aren't.
31:35And let's let's let's bring that up.
31:37Let's actually talk about that.
31:38So, when you when when we look at this,
31:39so you know, an average student in the
31:42class or our median student in the class
31:44is doing spectacularly well. However, we
31:46have a long tail. Every time we teach
31:5061C, there is a tail. There are people
31:52who are one week, two weeks, three weeks
31:55behind. And that is evident this
31:57semester. I mean, this semester is more
31:59difficult than any other semester that
32:01we have and we do want to help. So, we
32:03don't want to the way how we designed
32:05the following
32:0761C relief package. Here we go. The
32:10relief package. I mean, let me let me
32:12let me queue it up so you can see.
32:13Exactly.
32:14To to relief package
32:17Go ahead. Go ahead. Help those that are
32:19not doing you know, not doing well to
32:21give them some relief to to be able to
32:23catch up.
32:24While providing equal opportunity to
32:26those that are doing great to continue
32:27doing great.
32:29So, let's get into it. We have a
32:3010-point relief package. 10-point relief
32:33package. And we want to spend some time
32:34we spent 20 minutes talking about this
32:36relief package to help everybody. By the
32:37way, it doesn't help to help the you
32:39know, just like normal relief packages
32:40don't just help the folks in the bottom.
32:41They help everybody in theory. So, this
32:43is going to help everybody in theory but
32:44particularly we're focusing this on
32:46students who are struggling.
32:48Yeah, exactly. So, the first thing so
32:51what they wanted to what also we wanted
32:52to say, we designed this course to be a
32:55little bit heavier in in front. And this
32:57naturally there was a
33:00drop off in in in the amount of things
33:02that needed to be done every week. But
33:04we are making that dramatically easier
33:06so people can catch up
33:09way before the semester is over.
33:11Um, the first thing that we had to deal
33:12with there was an issue with speech this
33:15free speech and deletion of some posts
33:17because things got a little bit out of
33:19control on Piazza. Um,
33:21what we did is um
33:24well, we talked to the TAs, we talked to
33:26the students and we do understand that
33:29there are some posts that were
33:30inappropriate but even though they were
33:33not completely they're a little bit
33:35heated, put it that way.
33:37They should not have been deleted. And
33:39we are not going to delete any posts
33:41going forward. We talked to our TAs
33:44about um
33:46trying to be better at moderating these
33:49um
33:50exchanges before they become heated. So,
33:52let's try to keep it that way. We also
33:55do want to ask students to work with us.
33:58Um, and we're going to be very
33:59supportive. We're going to work with our
34:01TAs to make sure that these discussions
34:04never get out of hand. Um, and we are
34:06going to be in here to step in like we
34:08have stepped in with the grading or
34:10regrading of the project 2B. There is no
34:14reason to for things to get out of
34:17control and get very heated and and
34:21to have inappropriate posts from
34:23anybody.
34:24Um, and so let's not do that. Let's if
34:27there is an issue, bring it to you know,
34:30put it on Piazza. If you don't get an
34:31appropriate response, you can ask
34:33politely to us. No need to to get very
34:35heated about it. We are very very very
34:38very reasonable about that. Um,
34:41And we appreciate the feedback. We also
34:43appreciate the feedback. I mean, every
34:44time you there was a concern that you
34:46know, if I if I'm critical the class,
34:48I'm going to have my EPA points docked.
34:50Actually, it's the opposite and we'll
34:52talk about this in a moment. We want to
34:53hear that feedback. That's why we give
34:54you so many surveys. We want to hear
34:56from you and we're going to institute
34:57some more policies. And part of these
34:59relief package we'll actually talk about
35:00how we're going to do more open channels
35:02of communication to help and get it you
35:03know, a pulse of how the students are
35:05doing. That's what the purpose of all
35:06these surveys was. Yeah.
35:08Exactly. So, let's move on. But again,
35:11we are going to sanction if somebody
35:12behaves inappropriately and then
35:15Don't don't don't test what's going to
35:17happen. Right. Yeah, I mean, we will
35:18bring in the center for student conduct.
35:19I mean, that's what happens, right? If
35:20you're if you're if you're violating you
35:23know, in some sense student conduct. If
35:24you're making the place uncomfortable
35:26for people to be there because you're
35:27doxing people, you're using explosives,
35:29you're doing other stuff in a space that
35:31is a learning space. It's just like if
35:33you were doing it on campus. We pull the
35:34center for student conduct in. We'll
35:36delete that post. So, we're not going to
35:37not delete posts. We'll delete that one.
35:39But when we delete a post, it's because
35:40we're pulling you into the center for
35:42student conduct for something that's way
35:43inappropriate. By the way, this happened
35:45once in my 20 years here. So, that's I
35:47don't think this is going to happen. And
35:49criticism is not that category.
35:51Criticism we welcome that. That's not
35:53what that is. So, we're talking about
35:54someone doing something way above where
35:56it's clear to everybody in class that
35:57was inappropriate. Don't do this, okay?
35:58So, we'll try to we'll try to be very
36:00clear when that happens.
36:01Yeah, and we are not going to delete any
36:03post. Back to you. Yeah. So, this is a
36:06big one. Um,
36:08we had a wonderful meeting with two
36:10students and I should three or four
36:11students
36:12over the weekend and I my eyes were open
36:15to how much folks are struggling. More
36:17so and in fact, it was really useful not
36:19just to hear some some other you know,
36:21another data point about that but to
36:23actually have a working path forward.
36:25And we were kind of throwing out what
36:26how would this work? What do you think?
36:27So, it was like a feedback space. It was
36:29really wonderful and I thank those
36:30students for being there.
36:31Um,
36:32one of the things that came up and a lot
36:33of things came out of this meeting. So,
36:34I thank for them is
36:36people they were very thankful you know,
36:38as a representation of face-to-face
36:39representation that we made project 2 a
36:42partner project. We mentioned before we
36:44had a fear that by doing that that could
36:46be a solution where one student one
36:48student doesn't do the work and let's
36:50just buys a coffee for the other friend
36:51they do the work. We we're just hoping
36:53that doesn't happen. We please you know,
36:54don't do that cuz you'll miss the
36:55opportunity to learn it yourself.
36:57We decided that it was better to do
36:59that. It was better you know, both sides
37:01all being equal it's better to have that
37:03person to ask questions to and support
37:05and and work together positively. We've
37:07decided to take every homework that was
37:09never a partner. There were never
37:10homework partners ever in 61C as far as
37:12I remember to make every homework going
37:15forward a partner if you want to. We
37:16won't force anybody there but we
37:18encourage everyone to have partnerships
37:20in your homeworks to do together, to
37:22talk about it. You know, you see the
37:24homework is like online. You click it
37:25kind of like it's a like a edX kind of
37:27thing. But it's a grade school. So, but
37:29we encourage you to do it live. Get on a
37:31Zoom call, talk your way through it,
37:33grab a piece of paper, work on the
37:34problems and learn together. So, we want
37:36people to learn together, form learning
37:37communities. So, boom. Number two is
37:39reduce your workload cuz you're working
37:40together in homeworks. Hopefully that's
37:42a big way to help with that. So,
37:44that's the second one.
37:45Over to you, Bora.
37:46The next one.
37:49homeworks nine and 11 are now
37:54not graded or actually you get 100%
37:56whether you you do them or not.
37:59That's supposed to free up some time. It
38:01is not a huge amount of time but it
38:02frees up a few hours in
38:05for each one of those. So, you can use
38:07that time to catch up. If you're caught
38:09up, please go ahead and do these
38:11homeworks.
38:12Next one is the same in the same spirit.
38:16So, uh
38:17on on three, the homeworks 9 and 11 are
38:19still out there. So, if you we encourage
38:21you all to do them but if you can't do
38:22them, no problem. Um, we'll we'll have
38:24you know, TAs are there to answer any
38:25questions about them but again, you get
38:26the free points. We're just giving free
38:27points away to support folks with
38:29workload.
38:30We had a really big conversation about
38:33labs. And labs could be oh, it's a half
38:35hour lab. It's an hour lab. It's a five
38:37hour lab. That's a really big deal.
38:39That's too many hours in that. So, we
38:40decide and by the way, the spirit of a
38:42lab check-in was it's the single place
38:45where every single student has to have a
38:46face-to-face high touch experience with
38:48a staff member in 61C. We spent all our
38:51TA budget in that really.
38:53And the idea you know, face-to-face it
38:54would be a lab and people you know, our
38:56our AIs and lab TAs are all walking
38:58around answering questions. It's a very
39:00healthy Q&A. You raise your hand. Two
39:02seconds later people come over and swarm
39:03and they answer questions. It's really
39:04healthy. You get a check-in and we
39:06verify you know it.
39:08We decided to have these lab check-ins
39:10here. And the idea was
39:13for every 15 minutes or so,
39:15a certain amount of time maybe three or
39:16four or five minutes is a health
39:18check-in. Social emotional health. How
39:20are you doing? How are you doing with
39:21your life? How are you doing with your
39:22health? How are you doing with 61C? Are
39:24you
39:24are you caught up? Are you confused
39:26about this or this? It's like a
39:27check-in. And then we start to the
39:29technical Q&A of like well, what happens
39:31here? Well, what if you do this and then
39:32what if you Okay, so that's the last
39:34part of it.
39:35And we were hearing this isn't working.
39:37We were hearing this isn't working. A,
39:38there's no check-in. There's no health
39:39check-in. And B, there's almost no time
39:40to do that. And students who are acing
39:43it and don't need that support stop by
39:45you know stop using up TA support to
39:47prove that the student knows it when
39:48they already know it.
39:49So, we decided that this is really a
39:51this is really an experiment. We've
39:52never done this before. We're trying
39:53this.
39:54We believe it's more important to do the
39:56health check-in than anything. We want
39:58to know how are you doing to have a
40:00chance for every student to have a very
40:02small meeting, you and your partner,
40:04with a TA with a staff member. So, what
40:06we're going to do is we're going to
40:07change the lab check-ins.
40:10You're still going to get the By the
40:11way, for the homeworks we're just going
40:12to be free points. That's how we deal
40:13with that one. But for the labs, what
40:15we're going to do is
40:16just show up.
40:18Talk to your TA. Tell them how you're
40:20doing. Talk about how your class is
40:21going. And then ask The rest of it is
40:24yours. Technical check-in driven by
40:26student. We encourage everyone to keep
40:28doing the labs. Do the lab. Then ask the
40:31questions of the lab of the TA. So,
40:33rather than the TA probing you, you are
40:35probing the TA. You know, I was stuck
40:36over here. What would happen in this
40:37case? And the TA is helpful, happy to
40:39help. So, it's driven by the students.
40:41It's whatever you need. If you need to
40:42just talk and vent, use the time to talk
40:45and vent. You will get points. So, come
40:47to those sessions to have a check-in.
40:49You will get points for showing up.
40:51That's the idea. Because of this, now
40:53we're not allowing any more late
40:54check-offs because there's no more
40:55content in those things. It's about a
40:57health check-in. So, it's not like,
40:58"Well, I can just show up to you know at
41:00the end and do this all No, you have to
41:02every week check-in and you get the free
41:03points if you just show up and have a
41:05conversation with our TAs. But again, if
41:07you want to make them technical, make it
41:08technical. Use the whole time to get
41:09support. One-on-one support. You want to
41:11use the whole time talking about your
41:12project your bug your bug you got? Use
41:14that time to talk about the bug. The
41:15point is it's for you.
41:16We're still testing the material. So,
41:18it's still out there. But if you if
41:19you're behind and you're oh I can't
41:21breathe. I can't do this and I'm lab is
41:23So, fine. Do it at the end. Do it during
41:25RRR week if you need to. But it lets be
41:27flexible for yourself. That's the idea.
41:29Okay?
41:31All right.
41:32On 0.5
41:34on this project 2B, we already said put
41:37a lot of effort
41:38in in re-grading it and we reshape the
41:42distribution by assigning more partial
41:45credits.
41:47Keep in mind that register calling
41:49conventions are really really important
41:51and even though if you do make a
41:54function that works on its own, if it
41:56cannot be called by something else, then
41:57you have a um
41:59a problem. I mean, it's going to be
42:01actually a nasty ugly problem that
42:03you're going to have a real hard time uh
42:05debugging
42:07in your practice as as a practicing
42:09engineer.
42:11Um so, hopefully this was a good lesson
42:13but it was an expensive lesson. Again,
42:15in the grand scheme of things in the
42:17number of
42:18uh points that
42:20people have lost,
42:22it's not a lot but it was still felt
42:25like a lot. So, we went and reshaped
42:28this distribution. So, hopefully people
42:30are feeling better about that about the
42:32number of points that they've lost. So,
42:33it's a just a small penalty on the
42:36calling convention issues.
42:39Um
42:40hopefully people are
42:43feeling a lot better about that now.
42:46And by the way, we thank every single
42:48student who brought it up to our
42:49attention that this was a little unfair
42:51and I got some emails and I want to
42:52thank you all. And by the way, you're
42:54going to earn EPA points for being able
42:55to speak up. That's EPA is effort,
42:58participation, altruism. I count
42:59participation altruism as what you By
43:01speaking up for the class, I speak up on
43:03behalf of the trees. That's the Lorax. I
43:05speak on behalf of the students. That is
43:07what EPA is about, being a good student.
43:09So, bringing that up, talk to us,
43:11improve the health of the class. EPA
43:12points for people who brought it up to
43:13me. Everybody send me a mail gets EPA
43:15points for doing that. Thank you for
43:16that.
43:17Okay, relief package. Point six.
43:21Again, this is I just got to I just got
43:22to mention this a second ago.
43:24There's And by the way, this number 14
43:2615 don't worry about that. That's our
43:28internal numbers for that. So, ignore
43:29what that means. There's never points
43:30off for complaining to us about feedback
43:32again. We're going to In fact, we're
43:34going to include students in our staff
43:37meeting. We're going to have a special
43:38It's never happened before. All these
43:39things are brand new in this case. So,
43:41we're going to have a couple of students
43:42as representative of the class and I'm
43:45going to ask the students to actually
43:46coordinate yourself and those folks will
43:48attend our staff meeting and give us a
43:50pulse. So, rather than hearing waiting
43:52till this meeting which is so great over
43:53the weekend, it was week nine I heard
43:54about these things. I want to hear them
43:56every week. So, we're going to have a
43:57chance every week to hear from the
43:58students and really support them and say
44:00like, "How did this go? How does our
44:01relief package going? And should we make
44:03other changes?" Well, that's you know,
44:04we have the power of the of the of the
44:06control of things. Should we drop
44:07something else? Do we need to drop
44:08another homework? How's the how's the
44:10work there? We want to support students
44:12in that way.
44:13Bora.
44:14Yeah,
44:15the other thing So, what we are
44:16realizing from Piazza posts and also
44:19from the quiet probably not majority
44:22minority in this case that there are
44:24students who are behind by a week, by
44:27two weeks and there is a huge
44:29appreciation for
44:32a session a last session that
44:35Caroline is doing on Fridays. We are
44:37going to add one more that is probably
44:40Apps going to trail Caroline a bit.
44:42We are trying to find a resource for
44:44that but I think we will get one so
44:46we'll have another
44:48catch-up so that you don't have to have
44:50a choice or do am I you know going to
44:52study this on my own or I'm going to try
44:54to
44:56to to attend the the current sessions
44:59and try to somehow you know go through
45:01this. We'll open up some time hopefully
45:04with a little bit reduced workload that
45:06you'll or significantly reduced workload
45:08with the homeworks and these catch-up
45:11sessions, people will be able to put
45:13time into that to catch up. Um such that
45:17we will have everybody
45:19pass through the gates through the
45:21finish line together on the same time in
45:24the end.
45:25Next point.
45:27Um
45:28I again this is a under the theme of
45:30support for struggling students. The
45:31folks who are not in that median of a
45:33hundred for most of those folks. We're
45:35going to try to do two things. One is
45:38the goal of moving to Discord was to
45:41simulate what happens when you're in a
45:43normal face-to-face line for office
45:45hours. There's an office hours queue for
45:46project office hours and you need to get
45:48you got a bug to get fixed. And normally
45:50there's a line outside the door and then
45:52students are sitting with their laptop
45:53and they're helping each other. They're
45:54talking continuously and often people
45:56will get their bugs solved and they
45:57leave before they even get to the TA.
46:00The purpose of Discord was to do that
46:01and I'm learning from students that
46:02they're not doing that. So, we idea I I
46:05don't know how to support that other
46:06than to say, "That's the place. While
46:08you're waiting, don't wait for six
46:09hours. Wait and be talking about things.
46:12We love that. Help each other. Don't
46:13give the answer away but help people
46:15debug. You know, how help people through
46:16the process of that." So, what we want
46:18to do is we want to create some videos
46:20and I'm going to ask my people who are
46:21the most you know Discord savvy to
46:23create some videos of kind of a best
46:25practice of how to do how to use Discord
46:27to get those questions answered, how to
46:28join, how to do this, and how to be
46:30while you're waiting be connected with
46:31other people.
46:32In addition, we created a help video for
46:34the project kind of like a here's how to
46:35think and we actually talked about it
46:37during this live session a couple of
46:38weeks ago. So, we're going to try to do
46:39the same thing where we have help videos
46:41to prepare you for project four which is
46:43a fun project but it's a big one. So, we
46:44want to make sure you're ready for that.
46:45So, help videos will kind of get Make
46:47sure you pay attention to this. That'll
46:48be in the help video.
46:49Bora.
46:51Yeah, um
46:52the the the other thing is
46:55This is also related to that. In normal
46:58semesters when we are in person, people
47:00have students have time to develop
47:03relationship with their TAs. TAs get to
47:05know them. They know where they are,
47:06what are their strong points, what are
47:07their weak points. Um
47:10And but we have a bit lost that this
47:13semester. So, what we do want everybody
47:16to sign up for the same
47:19slot with the same TA for the office
47:22hours
47:23such that they can they get to know you.
47:27One of the challenges that we see here
47:29is that
47:30TAs are not exactly getting a handle on
47:33knowing exactly you know who you are. We
47:36do want you to to um to develop a
47:39relationship with TAs. TAs have a a
47:41chance to to keep their hands on your
47:44pulse and know how you're doing if
47:45you're doing well or you you're suddenly
47:47struggling after a week. They can tell
47:49us that and we can
47:51do something. We can intervene. So,
47:53please please do that and we do want to
47:56to have that
47:57um
47:58uh in there. We want you to be
48:00participating with the TAs.
48:04Great. And finally,
48:06this is the biggest one. I saved this
48:07for the last one. This is the biggest
48:09one. I I We talked to Bora about it and
48:11he was on board so I wanted to mention
48:12this. This is a really big deal.
48:14I want to have a safety net for all of
48:16you. I want to make sure that none of
48:17you feel like it's just I can't deal
48:19with it. It's just not worth it. This is
48:21this this this course is is too much.
48:23I'm going to offer right now
48:26to say that I will not fail anybody this
48:28semester. That's not to say I'll give
48:30you all passes. What it says is you say,
48:33"Well, the course got off the rails at
48:35some point for me and I couldn't keep up
48:36with it."
48:37All you need to do is mark that date for
48:39me. I'll have a form that people can
48:40fill in and I will give you an
48:42incomplete. I've never offered this
48:43before. I offered to CS 10 in the spring
48:45and it worked. Never offered at this
48:46scale to 61C before. What it means is
48:49you'll get an incomplete for the grade
48:51if if choose this if if the if it's just
48:53not going well for you and you say,
48:55"Dan, yes, I even did get it to week 15
48:58but it really I was feeling not doing
48:59well since week 10." Well, go back and
49:01we'll backdate your incomplete to week
49:0310 and your grades after week 10 won't
49:04be counted and you'll be able to redo
49:06those again.
49:07So, it's a chance for you if you were
49:08going to end up in the failing area to
49:11to pressure you. You can always reach
49:12out and say I just say well incomplete.
49:13I'll say yes to every incomplete request
49:14period. And then everybody who would
49:16have normally failed, which is a small
49:18percentage maybe like 5% or something,
49:19very small, but I hate it anyway, I'm
49:21not going to give them Fs. I will or
49:23NsPs. I will give them Is.
49:25And then you have next semester or the
49:28summer and I'll always extend these
49:29incompletes as long as they take. Okay?
49:31So, no fails this semester, only
49:33incomplete for those folks who would
49:34have failed and reach out if you need
49:36some help. That's out there as a safety
49:37net. So, don't think it's the end of the
49:39world. We're here for you.
49:41But do reach out.
49:42You know, that we we have seen people
49:44who have been dropping the class and I
49:47said and intervene, "Hey, I mean, you
49:49don't have to drop it. We can give you
49:50an incomplete. I mean, you you you don't
49:52have to redo the whole thing. You Yeah,
49:53you put work in. Keep keep keep the work
49:55that you've kept. You don't don't throw
49:56that all away. Exactly. If you drop it,
49:58you don't get to keep the work you've
49:59done. You get an incomplete, you do.
50:00This is a really big deal. I don't know
50:01about any other class. Tell me if
50:02another class CS class is doing this.
50:03I'm doing this. I did it in the spring
50:05and it worked. I'm doing it again here
50:06to support everybody at the bottom you
50:08know the at the bottom level here.
50:10I mean, these are extraordinary
50:11circumstances, right? I mean, that's the
50:13justification for an incomplete. It's
50:15it's default this semester.
50:16Out of your control and COVID is out of
50:18your control and this whole situation
50:19out of your control, right? This whole
50:21system is out of your control. All
50:22right. We got some more things quick.
50:24It's at 5:00, but I want to make sure I
50:25cover these things. Yeah. Um 53 people
50:27took the quest a revision of the quest
50:30last Sunday. We did not shut down
50:32Piazza. There was so much things
50:33happening this weekend. We didn't want
50:34to shut that down, but we let it happen.
50:35People were very very good. We let every
50:37student know you need to ask a private
50:39question on Piazza. 100% of you did the
50:41right thing. So, thank you for all, you
50:43know, walking the line and keeping that
50:44cuz otherwise it'll be an issue where
50:45you're giving the answer away and all of
50:46a sudden students are seeing that. So,
50:47we didn't shut it down. People were very
50:49respectful and did this. We will shut
50:51Piazza down for the big midterm cuz
50:52that's for everybody, but we didn't want
50:53to shut it down for this small group of
50:54students and it was similar to the quest
50:56in terms of style. Not everybody had to
50:58do all the all the parts only the parts
50:59that we didn't have a video for you for.
51:01So, those folks had to do them again.
51:03We're going to have to write custom auto
51:04graders. You saw how long the regrade
51:06took. It took 3 weeks for that. Now it's
51:07a smaller load because it's mostly the
51:09same as those auto graders, but but we
51:10have to write them now and so that we
51:12hopefully you'll hear back to us this
51:13week about your score.
51:16Finally, the big deal. Uh we've got a
51:19midterm. The midterm is scheduled. You
51:20know about this. This is last week's
51:21information. It's this coming Sunday and
51:23Monday. The material covers we Bora has
51:26mentioned is up to two up to and
51:27including pipelining which is today's
51:29lecture. There's going to be exactly
51:30five questions. The topics are on the
51:32bottom. You got a quest redo which is
51:33coding in C. You got to float question
51:35very small lightweight question there.
51:37RISC-V assembly with which has some
51:39coding so practice that. Data path
51:41control and pipeline and an SDS logic.
51:44We will give you tests when appropriate.
51:46There's one small part where I can't
51:47give the test cuz really the test is
51:48about can you think about how to test
51:49it, but I'll give you a structure for
51:51the test around it. So, it'll be
51:52structure for two tests there ready to
51:54go there. Um
51:56time limit of roughly an hour per
51:57question. We're going to share our TA
51:59test times. Our TAs are beta testing it
52:00now or they'll be testing I think this
52:02evening they start and all those beta
52:04testing times we're going to share with
52:05you. We'll say a TA took this many
52:07minutes, this many minutes, and this
52:08many minutes the range and then we're
52:10going to take that time and multiply it
52:13by a factor like three or four and
52:15that's how much time we'll give you. So,
52:16we'll take the maximum time any TA took
52:18to take it, multiply it by a factor, and
52:20that's what you should take, okay? You
52:22should expect to take at least 2 hours.
52:24I would say two to three is the safe
52:25space of reserving yourself, but leave
52:27yourself more of that cuz you know, if
52:29you if it's an hour per question, that's
52:30at least maximum five, but we're not
52:32going to say it's 48 hours, but we're
52:33going to say we think we can get
52:35everybody done at like a multiplying
52:36factor of about four times the longest
52:39TA time ever. So, that's what we're
52:41going to do. We didn't do that at the
52:42quest. We're going to do that here to
52:43give you enough time, but also constrain
52:44it so it's not a 48-hour you're staying
52:46up the whole time, okay? Lots of
52:47questions in Q&A. I'm going to stay
52:48until those get all get answered. So,
52:50let's let's hear.
52:52I Yeah, I I did some how to review
52:54quest. So, they're asking for quest
52:57when can we you know, now since the
52:59retake is done
53:01Right. when can we put
53:03Well, I think we were already shared the
53:04answers. I believe we already shared the
53:05answers for the first version of the
53:07quest and if not we I don't have a
53:08problem with giving those giving that
53:09away. I think that's a a little piece of
53:11code and we can give that away. So,
53:12we'll do that um if that hasn't done
53:14already.
53:15So, so I'll get that answered live.
53:18Um
53:20Uh there's an interesting question
53:22that um has come up in our conversation
53:25which we had it's not part of our
53:2610-point relief package which is could
53:28you use your slip days to resubmit
53:31project 2B? Make the fix. You know, our
53:33hidden tests are still hidden. So, what
53:35if you could use a slip day to actually
53:38fix your code and then resubmit for an
53:40updated grade? I'm actually okay with
53:41that. I want to check with Bora. I
53:43didn't get a chance to talk about this,
53:44but it happened while we were giving
53:45live the Yeah, we'll just discuss it. We
53:47might have good news for you next week
53:49if we decide that that's okay. So, we're
53:50we're we'll we'll think about it, but
53:51that would be amazing to be able to play
53:53with that. I I love it. So, I'll argue
53:55in your support for that, but other
53:56people may say, "Well, it's not fair."
53:58Blah blah blah. So, let's hear all
53:59perspectives before we let you do that
54:01one. Okay. Um
54:04So, Will the check-in
54:06obviously be effective for lab six if we
54:09were planning to get checked off late?
54:13I think I think I think the check-in
54:14policy is now. If you have if you have
54:16content, if you're ready to be checked
54:18off, go ask those technical questions. I
54:19don't see why we couldn't start that as
54:21soon as we can, to be honest. I think
54:23meaning you can get full points
54:25that's Oh, for last week's lab. No, you
54:27should you should finish last week's
54:29lab. It That's that's the starts from
54:30now forward. Yeah. Starting making it
54:32forward. So, checking off late you can
54:34still get to check off late, but just
54:35use the time during your session to
54:37check off last week's one going forward.
54:40Exactly.
54:41Um Can we check in for more than one day
54:45a week and when do revamp check-ins
54:47start? Revamp check-ins start next you
54:50know, they can start this week. We don't
54:52have a lab this week. Yeah, yeah.
54:56Um I I
54:58I I I mean, we have to look to see if if
55:00more you know, if you wanted to have 2
55:02hours with our TAs. If we have the time
55:03for it, I don't I don't know. I think we
55:04have to continue to make sure that every
55:06TA can cover every student. So, I'm not
55:08sure I'm going to guarantee that, but I
55:09think it should lighten the load in
55:11terms of meaning some people may be
55:12using their check-in to get their
55:14question answered for their project and
55:15all of a sudden now project office hours
55:17are lighter and normal office hours are
55:18lighter. So, I think we're going to
55:20lighten the load when people are making
55:22use of these check-ins. So, just give
55:23more face-to-face time with your TAs
55:25that isn't about the lab. You say, "You
55:26know what? I learned the lab during RR
55:27week." We encourage you to learn it
55:29because sometimes the projects need it,
55:31but I'll learn that particular lab that
55:33isn't on a project later. I'm going to
55:35use the time now when I really need it.
55:36That's fine. I think that'd be great and
55:37that might lighten the load on other
55:38times. So, I would say I would say right
55:41now no, one check-in a week for you, but
55:43come to other There's a ton of office
55:45hours all throughout the week. Hit those
55:47and and do that one if you can.
55:50Um
55:51There's
55:53uh let's see. 2B score was not supposed
55:56to decrease. That's a that's a question
55:57about this. 2B was not supposed to
55:58decrease after these changes. That was
56:00actually one of the things we talked
56:01about, but it it may have. We want to
56:03I mean, it could happen. I mean, in in
56:06some corner cases it could have
56:07happened. Please let us know. We'll give
56:08you the higher of two.
56:10Yeah, I think that's right. I think you
56:11get the higher of the two options. So,
56:12there we go. There's the max of either
56:14the first the old one or the new one. We
56:15can easily do that in software. So,
56:16let's do that one. I think that's fair.
56:18Yeah. Um
56:20Uh I believe I I don't know. APA points
56:24are just a small fraction. They're not a
56:25big deal, but a lot of people are asking
56:27about APA. Um
56:28I don't know if you comment anonymously
56:32whether at the end of the year we
56:33download the Piazza participation and we
56:36use that to inform not to completely
56:37override, but to inform our APA and I
56:40don't know if when you comment
56:41anonymously whether it gives you credit
56:44for that. It's hard to test. You have to
56:45like test it with a with a blank Piazza
56:46and make an anonymous comment and see if
56:48you get a note for it. Um I know who
56:50Bora and I can see who wrote the
56:51comment. So, the staff can see. I don't
56:53know if the software sees. I can ask
56:55that question. I'll try to look into
56:56that. I'll just say that. I don't I
56:57don't know the answer to that one. I
56:58would like it. I would like to be able
56:59to be anonymous so that you don't get
57:01ire from your neighbors, but you should
57:02be able to get a credit I would say for
57:04at least participating in that way. Um
57:08Go ahead. You want Bora to grab one?
57:09Hey, we we are at 5:06. So, we we pray
57:13for
57:15Uh okay, I'll do easy ones. Midterm is
57:17open notes just like the quest. The
57:19midterm is almost identical. Midterm yes
57:21on PrairieLearn. Yes, almost identical
57:23there. That's easy. Um
57:26Uh
57:28What Yeah, what would the quest coding
57:29contain number representation questions?
57:31The answer is everything. Everything up
57:32till now. So, the answer is everything's
57:34live. Who knows what's going to be on
57:36it, but yes, it would potentially
57:37contain number representation questions.
57:38Know that stuff, too. Important. Um How
57:41should we reach out? Private
57:43Piazza you know, that's that's the best.
57:46Uh private Piazza is one way
57:48and but the second way I think is good
57:50we we need it.
57:51Yeah, for me email is not the best way
57:53cuz I have too many, but I would say
57:55pri- Reaching out the best way to do
57:56this is from at least my point of view,
57:58talk to the student group that's going
58:00to be the representative students group
58:01talking to us once a week and that'll be
58:03a chance for you to join that group or
58:04at least get your voices shared. We want
58:07to have a small group of students
58:08talking to us representing the students.
58:09So, talk to your representative who
58:11comes to meet with us once a week. So,
58:12we'll try to organize that and
58:13coordinate that and I'll make with the
58:15student who volunteered to kind of maybe
58:16be part of that. So, that's a great way
58:18to do that. Um boy, a lot of questions.
58:21You know, labs are there's an auto
58:23grader that is going to you know, you
58:26there is still a lab auto grader that
58:28you but it just gives you 100%.
58:32So, that's Well, I mean, at the end of
58:34the day the grade Now, at the end of the
58:36day the points in our in the in the
58:37grade file will be 100 if if you show
58:39up, but the auto grader might give you
58:41feedback. So, play with the auto graders
58:42if they're a lab auto graders to learn
58:44whether you got it right or wrong. You
58:45got to know whether you got it right or
58:46wrong or not, but at the end of the day
58:47if you showed up live, you get the
58:48points in the in the grade book. Um can
58:50I use a C reference guide in the midterm
58:51for syntax purposes? Yes, download any
58:54reference guide you have in advance. You
58:55just can't search for it live. You can't
58:57say, "Oh, the problem asked you for a
58:59make this kind of shapey thing a snakey
59:00thing." You can't say, "Google snakey
59:02thing." You can't do that, but if you
59:04download anything in advance, you can
59:05have it. So, a great idea. Get any any C
59:08guide you need to do that. And have as
59:10well as have your C have your book. It's
59:11open book everything. You just can't
59:12search anything live so you don't do
59:14that. That's perfect. Um how crowded are
59:16professors' office hours? That's easy.
59:17Bora, how crowded are your office hours?
59:20Um they were not in the past.
59:24I had about 10 people on Friday, but
59:26they did my office hour instead of for 1
59:28hour I spent 2 and 1/2 hours on Friday
59:31and I intend to run the two hour office
59:33hour again on Friday.
59:35Um my office hours I have two of them. I
59:37have a one-on-one for particular special
59:39problems for students need to talk about
59:40maybe a 47C thing. But I have my general
59:42office hours are mostly what I do
59:43normally in the in the face-to-face
59:45here, which is I go over midterm
59:47questions and I try to help people with
59:48exams. So much more But I also have you
59:50know we talk about anything that's
59:51coming up, but I mostly go over content
59:53in mind. So I'm happy to have both. The
59:55first one the content is I don't know
59:57about what we had the people have been
59:59there for that group about 10, which And
1:00:00so I welcome my office hours this same
1:00:02time on Wednesday is my content office
1:00:04hours. And one-on-one you can make a
1:00:05one-on-one appointment if you have
1:00:06something special. But again, many
1:00:07one-on-ones can come through that
1:00:08representative thing. Um Ryota asked
1:00:10some very good questions. Go ahead or
1:00:12please. Yeah, we have we have two two
1:00:14types of office hours. If you have
1:00:15something that is bothering you, come to
1:00:17the individual and there is no one else.
1:00:18You just come and and talk to us for 10
1:00:20minutes. If if there is nobody in the
1:00:22queue with me and typically after 5:20,
1:00:255:25 there is nobody. Somebody shows up
1:00:27at Oh, sorry, 4:20. You know, usually
1:00:30there is a few people who are queued up
1:00:31with immediate issues.
1:00:33Um I clear that by five 4:25 4:25 and
1:00:36then somebody shows up at 4:50. Uh so
1:00:39you you you can grab that time. Um
1:00:41again, if I advertise that it's going to
1:00:43get crowded. So those are my Wednesdays
1:00:45ones. Then the Friday ones are are
1:00:47content ones and I go over the issues of
1:00:49whatever the the issues are and just
1:00:52come and and ask questions. Last time it
1:00:54we went for two and a half hours until
1:00:56we were done with the questions.
1:00:58I can't guarantee that I can do every
1:01:00Friday uh two and a half office hour or
1:01:02half an half hour office hours, but I'll
1:01:04try.
1:01:05Right. Perfect. I've got two easy
1:01:07questions I can hit. Can you download
1:01:09websites for reference? The answer is
1:01:10yes. Pre-download websites. The point is
1:01:13you just can't search for it. When I
1:01:14give you a problem about the sneaky
1:01:15thing, you can't search for the sneaky
1:01:16thing. But if you want to download all
1:01:18the questions you have, download them
1:01:19all, that's perfect. And that's
1:01:20Monisha's question, good question. And
1:01:22the second is um can you use local
1:01:24computer search? You know, I built it to
1:01:25this big directory and I'm searching
1:01:27through my directories for where I put
1:01:29that silly thing. Where did I put that
1:01:30website that had about to do that? You
1:01:32can do that. You just can't ask Google.
1:01:33You can't ask Bing. You can't ask the
1:01:34web. So yes, 100% local computer search,
1:01:37very very good. That's in the spirit of
1:01:39that exam. Um
1:01:42You want to jump on one more or maybe
1:01:43more? Those are the two easy ones that I
1:01:44got here.
1:01:46Um How do we sign up to be
1:01:48the representative? I think we're going
1:01:49to ask somebody
1:01:51to be a representative.
1:01:52Yeah, we we might be to maybe to a
1:01:53Piazza page who wants to be and we'll
1:01:55have a couple folks. And then what that
1:01:56Here's the point. You'll coordinate. The
1:01:58representatives will then coordinate a
1:02:00student meeting and they'll check in
1:02:02what do we want to share with this? And
1:02:03then those representatives will come to
1:02:05our meeting. We wanted to have a couple
1:02:06people who are really on point with the
1:02:07bulleted set of things. But we want to
1:02:09have that conversation outside there. So
1:02:11we'll try to let it be self-organizing.
1:02:13Ryota, you want to organize it? Organize
1:02:14it yourself. And then get together to
1:02:16get group folks together and then you
1:02:18and the person who I met with earlier
1:02:20will then be representing. So we'll
1:02:21figure out who those representatives
1:02:23are. You guys can choose among yourself
1:02:24who wants to be the representative to
1:02:25come to talk to us, okay? Love that.
1:02:27Thank you for for organizing that or
1:02:29thinking about that. We'll set up a
1:02:30Piazza thread for that. Good stuff.
1:02:32Um
1:02:34Let's see. Uh
1:02:36saving slip days to the end, that's been
1:02:38an interesting conversation, a design
1:02:39decision. I think I've never allowed
1:02:41slip days to count for anything. If you
1:02:43don't use any slip days, you can use
1:02:44them for points. We've never done that
1:02:46in the past. I prefer you use your slip
1:02:48days to just give yourself more The
1:02:49point is it's flexibility. If you If you
1:02:51If let's say you know if you get sick or
1:02:53something, you get a you get the
1:02:54sniffles and you need to take a break or
1:02:55you're overloaded. It's about that
1:02:56space. And here's the problem. The
1:02:58reason I don't like people saving it at
1:03:00the end is if you get sick, I want you
1:03:02to use them. I want you I don't want
1:03:03them to become currency to become free
1:03:05points later. Because then when you do
1:03:07get sick, you'll like you'll just bury
1:03:09through it. I don't want that. I want
1:03:10you to be taking your breath, rest, you
1:03:12know,
1:03:13drink if you know fluids and rest and
1:03:16take the break when you need it. So I
1:03:17actually prefer them to be worth nothing
1:03:19at the end. So that's my policy in the
1:03:21past and I want to stay with that. Which
1:03:22means you can still use them up for the
1:03:24last project, that's fine. So save them
1:03:25for the last project. If you don't use
1:03:26them for the last project, you won't use
1:03:27them at all. But I don't like the idea
1:03:28of saving them up for something else cuz
1:03:30I want you to be using them if you need
1:03:31them. So don't I I I think the answer is
1:03:33I'm not going to let you do use slip
1:03:35days for points. I It's an interesting
1:03:37idea.
1:03:37There is no class.
1:03:39There is no class in this slot, but that
1:03:41Oh yeah, you know, I think we're good
1:03:42with the with with
1:03:43class to overlapping. That's yeah. Um
1:03:47Uh more slip days? I think for now no. I
1:03:50think we always have the possibility to
1:03:51give you more later. That was just a
1:03:53question here, but we gave you three
1:03:54more and now you have six. Six is a lot.
1:03:56We've never had more than three in the
1:03:57history of 61C and now we're doubling
1:03:59it. So I think that's enough to be able
1:04:01to be flexible, especially with no more
1:04:02labs. You have to do project, you know,
1:04:05teamwork, homeworks. I think that's a
1:04:06little bit, you know, the 10 point uh
1:04:09relief package wasn't up. So we'll think
1:04:10about it. It's in our back pocket to
1:04:11give you more, but for now I think the
1:04:13answer is no for that. We'll see if we
1:04:14can do that. Okay?
1:04:17Um
1:04:20So the lab points do they only come from
1:04:22health check-offs plus technical
1:04:23questions if you want them? Uh standard
1:04:25autograder check-off component. No,
1:04:26there's no standard autograder. There's
1:04:28no check-off component. You just showing
1:04:30up and using the time in whatever way
1:04:32you want the time to be used gives you
1:04:34the points. That's the It's basically
1:04:36free points now. The lab Usually lab
1:04:38people be most people ace all the lab
1:04:39points. But now if you want to ask the
1:04:41questions about that, it's on you now.
1:04:43It's actually more ownership of you to
1:04:44understand that material, less about we
1:04:46need to push you because you know, we're
1:04:48the big brother and making sure you get
1:04:49No, if you need the questions asked at
1:04:51the end of the day, we've got an exam
1:04:52that's going to be testing that
1:04:53material. If you want to learn it at the
1:04:55just-in-time or you want to learn it
1:04:56during the lab, I would learn it when
1:04:58it's happening. Don't just fall behind
1:04:59it. So I encourage you to do it then,
1:05:00which I which I love. So that's great.
1:05:03Um Here is an important one. Will the
1:05:04midterm have similar proctoring to
1:05:07to the quest? Yes, it's identical.
1:05:09Identical. It's identical, which means
1:05:11plug in your phone. Oh my gosh. Plug in
1:05:14your phone. I can't have people say oh
1:05:15the phone went out. And turn off the
1:05:17notifications. We're going to have a
1:05:18whole checklist. We're actually editing
1:05:20a whole document with like turn off
1:05:21notifications, plug in your phone, turn
1:05:23Practice and then try calling yourself.
1:05:26Try getting a notification. Try to
1:05:27create a notification to see if it stops
1:05:28the Zoom recording. And by the way, of
1:05:30the two options, I actually found it
1:05:32much easier for students who created a
1:05:34time lapse. So there's any way for you
1:05:36to download software to create a time
1:05:37lapse, I encourage you to do that one
1:05:39cuz I actually found that that was much
1:05:40easier to see the screen and much easier
1:05:42because a lot of Zoom crashed for a lot
1:05:43of people, but the time lapse didn't. I
1:05:45had far fewer problems in the time lapse
1:05:47than I had with Zoom. So think about the
1:05:48time lapse and get your phone to do that
1:05:49if you can or download a free app or
1:05:51something like that. Great job. Um
1:05:54Five question means five midterm
1:05:57sections like the quest had three sec
1:05:58Exactly. Five questions We already
1:06:00talked about them. That's on the slide.
1:06:01Sorry here. Look, it's right here. Here
1:06:03are the five in my fingers. See my
1:06:04fingers? Look. Here's my fingers. Those
1:06:06are the five questions. Each of those
1:06:08Sorry, five sections. Five sections. I
1:06:12should I should edit this. It's It's
1:06:13quest five sections. Each section might
1:06:15have multiple questions. Okay? I'm sorry
1:06:17for not for not being clear on that.
1:06:19a sec you know new questions per
1:06:21section. Make sure you finish the whole
1:06:23exam. People get confused that. Don't
1:06:24don't do that, okay? So that's easy. Um
1:06:29There won't be a guide for writing
1:06:30tests. Essentially, we're not going to
1:06:32grade you on the test. We're just going
1:06:33to give you an easier framework to be
1:06:35able to have tests. Almost like for
1:06:36example, in your quest many of you had
1:06:38to do a split. You had a linked list of
1:06:40things you had to split it in some ways.
1:06:41There were different versions, but you
1:06:42split this, right? I didn't give you
1:06:44anything.
1:06:45Where now I would give you something
1:06:46that if you did it, it would split a
1:06:48simple thing. If you did it the right
1:06:49way, it would split a simple thing. It
1:06:51would test whether the split worked.
1:06:52We're going to give you that thing to
1:06:53test whether the split works. That's the
1:06:55idea. Or at least print it out. We'll
1:06:57have something I might even print a
1:06:58linked list. That will even let you see
1:07:00it whether it split works. So if you can
1:07:01split it and now I have two of them, I
1:07:02can print both of them. Did it work? So
1:07:04we'll give you that kind of framework to
1:07:05do that. It's not a formalized
1:07:06framework, but it'll be an easier way
1:07:08for you to test without you like start
1:07:09from zero to do that, okay? That's the
1:07:11And by the way, remember it's four times
1:07:12three or four times the time of my
1:07:14longest my slowest TA. So hopefully
1:07:17you'll have enough time. And they don't
1:07:18get it either. So we'll make sure that
1:07:19they can do it in a reasonable time and
1:07:20then multiply that by a factor to give
1:07:22you that. So that's good question. Okay.
1:07:24Keep going. We I'm still here. I'm still
1:07:26here.
1:07:27Um
1:07:29SSH'ing into to exams to something like
1:07:31Valgrind or CGDB? Yes, indeed. If you
1:07:33want to use If you want to SSH into the
1:07:36to the um
1:07:38to the hive, the hive If you want to
1:07:39code on the hive because the hive has
1:07:41all the software, please do. We We
1:07:42actually answered that on the quest.
1:07:44SSH'ing. So practice that. Practice
1:07:46Okay, everybody practice to make sure
1:07:48you know how to code in C, print stuff
1:07:50out, debug, use Valgrind, use memcheck,
1:07:52all those things. And then practice
1:07:55coding in RISC-V. If you haven't done
1:07:56that yet, make sure you do that, okay?
1:07:58And we're going to try to have a
1:08:00practice midterm that'll have a very
1:08:02simple question. Won't be the same
1:08:03difficulty. It'll be a question that'll
1:08:04be like do something, you know, uh
1:08:06do something but write I think I'm I'm
1:08:07suggesting to say write Fibonacci in
1:08:09RISC-V, which is something we do in like
1:08:11lab or homework or or discussion. And
1:08:13just say write Fibonacci in RISC-V and
1:08:15you'll be able to say okay, wow, protect
1:08:16this is actual exam. Could I write
1:08:17Fibonacci in RISC-V and print it and
1:08:19make sure it works? If you can do that,
1:08:20you're probably 90% of the way there in
1:08:22terms of getting the RISC-V stuff work.
1:08:23And the C is the same way for the quest.
1:08:25Good question.
1:08:26Good question. Just make sure that it's
1:08:28visible that that that
1:08:30you know, that that we can see what what
1:08:31you're using in in your recording.
1:08:33Exactly. Exactly.
1:08:35Um Um Next question. For time lapsing,
1:08:38are we making videos for each question?
1:08:40Yes. Five videos. Five videos, exactly
1:08:43right. Yep.
1:08:44Five sections, not questions. Each
1:08:45section. Yeah, each section. Sorry, we
1:08:47got to keep make be clear about this.
1:08:49And take a break between it, right?
1:08:50Don't don't go straight. A lot of people
1:08:51actually did it straight. It's I don't
1:08:53think it's healthy. Take a section,
1:08:55go take a drink, hit the restroom, have
1:08:57lunch, then do another section, jump
1:08:59back in. Some of them are really short.
1:09:00The float is going to have two question
1:09:02for the one section, two. I used to have
1:09:04four or five, two. So I'm light We're
1:09:06trying to remove the workload but still
1:09:08testing the important stuff. So we're
1:09:10going to try to have fewer things per
1:09:11there and so that'll be hopefully there.
1:09:13Um Venus is okay to use. So Venus
1:09:17Certainly yes. Venus is okay to use.
1:09:19That's how we encourage people to do it.
1:09:20If you're going to code RISC-V, test it,
1:09:22debug it in Venus. We're going to have a
1:09:23whole structure that still isn't written
1:09:24yet. We're going to try to make the
1:09:25whole structure there. So that's how you
1:09:27test good question viewer. You're doing
1:09:28really well with good questions today.
1:09:30Um
1:09:32Logisim questions. I don't think Bora,
1:09:33if you want to talk about whether we're
1:09:34going to ask them to use Logisim or not
1:09:36or what the plan is for that or whether
1:09:37they can use it. What do you think? We
1:09:38want to talk about where does Logisim
1:09:40fit into the to the piece of this? We
1:09:42want to We want to reel the question but
1:09:43but in terms of that piece of that tool,
1:09:45what How do you want to talk about that
1:09:46one? What do you think?
1:09:48I think we can say that we we were
1:09:50thinking about using Logisim but we are
1:09:52not going to um
1:09:56make you use Logisim.
1:09:58Right. So, there's no forced Logisim
1:10:00right now is the current plan and I
1:10:01think that that's not going to change.
1:10:02Um but I'll ask from the students' point
1:10:05of view, could they use Could they use
1:10:06Logisim to kind of help with that
1:10:07problem with that question?
1:10:08That's fine. Go for it.
1:10:09Okay. Okay, great. So, that's
1:10:113.1.
1:10:12Yeah. Okay, good stuff.
1:10:14Perfect. Perfect. Perfect. All right.
1:10:15So, answered live. Wonderful.
1:10:17Um
1:10:19I saw it. Can you have two videos? One
1:10:21for the physical workspace and the other
1:10:22for the screen. That's perfect. Yeah. If
1:10:23you want to have a lot of people do a
1:10:24lot of people do the screen video and a
1:10:26workspace, that's great. That helps us a
1:10:27lot so I can see. By the way, whatever
1:10:29you choose, I should be able to read
1:10:31what's on your screen. So, if you give
1:10:32me a video and we were really soft on
1:10:34this in the quest but we're going to up
1:10:35the bit level. I have to be able to see
1:10:36what's on your screen. So, if you can't
1:10:38See, if you're going to try to zoom call
1:10:39it and it's all fuzzy and I can't see it
1:10:41or the or the or the contrast is bad,
1:10:43you got to practice that and make sure
1:10:45we can read your screen. This is Now,
1:10:46this midterm is worth more than just a
1:10:48quick quest. We didn't care as much.
1:10:49It's only worth the 5%. We don't care.
1:10:50We'll trust you on that but we got to be
1:10:52able to see your screen. So, if you need
1:10:53to do a a time lapse sometime Time
1:10:56lapse, I can see everybody's screen.
1:10:57Everybody who does a time lapse, I can
1:10:58read their screen. I can read the
1:10:58smallest font.
1:11:00But the A lot of zooms, I couldn't read
1:11:02the screen. We're going to say you have
1:11:03to be able to give me a way for me to
1:11:05read the screen. So, if you're going to
1:11:06try to zoom, if you want to do that,
1:11:07that's fine but you got to make a screen
1:11:08movie, too. I got to be able to Whatever
1:11:10you do, I need to be able to see what's
1:11:12on your screen. That's that's the key.
1:11:13Okay? So, you can't just have a little
1:11:15the window and doing something inside
1:11:16and I can't see it. So, practice this.
1:11:19Adjust your light. Adjust the back
1:11:20light. Play Play with your phones. Play
1:11:22with all those things to be able to get
1:11:23that. But and and by the way, I still
1:11:25So, I love the idea of two videos to do
1:11:26that and we'll have a way for you to
1:11:27upload those two videos. I would just
1:11:28zip We'll probably have a one link and
1:11:30put them in a in a zip file and then put
1:11:32them there as one thing. That's a great
1:11:33question viewer. Again, you're doing
1:11:34really well for questions today.
1:11:36Um
1:11:38Let's see. Uh
1:11:42You don't need to submit labs to the
1:11:43autograder. Correct. The labs, you could
1:11:46submit to the autograder to get feedback
1:11:48for yourself but you don't have to.
1:11:49That's the idea. You will get full
1:11:50credit for that lab for showing up and
1:11:52talking to our TA and that can be all a
1:11:54technical conversation or a health
1:11:56check-in but you don't need to submit
1:11:58things in the future for all labs to the
1:12:00autograder. The autograder is there to
1:12:01help you but not going to be something
1:12:03we'll grade. Okay? The grades at the end
1:12:04of the day for the lab is not going to
1:12:05be part of the autograder. You can You
1:12:07can mess them all up and never get them
1:12:08right. You'll still get credit for your
1:12:10lab. Okay? It's more about checking in
1:12:12So, so but use the autograder. It is It
1:12:14is useful.
1:12:14Yes, definitely use it. Definitely use
1:12:16it. That's great feedback. Great
1:12:18feedback. Yeah, love that. Love that.
1:12:20Um
1:12:24Uh
1:12:25let's see.
1:12:28I guess there was a there was a there
1:12:29was from the chat uh this I I didn't see
1:12:32the chat here. They're saying Discord
1:12:33isn't working partly because the TAs
1:12:34aren't there enough for things TAs need
1:12:36to help with. I mean, the whole point of
1:12:37Discord is we have our TAs to help with
1:12:39questions. When you finally get to the
1:12:40TA, ask whatever question you want to
1:12:42have with a TA. But while you're
1:12:43waiting, use Discord effectively to ask
1:12:45all the questions that there. If you
1:12:46take the sum of knowledge of how to do
1:12:48stuff like set up set up a repo or do
1:12:50these things with the students, it way
1:12:52exceeds what TAs know. So, students can
1:12:54help each other in ways that can scale
1:12:56with students. TAs are a limited
1:12:58resource and they're overworked this
1:13:00semester. I'll be very honest. So, there
1:13:02might be a long line but help each other
1:13:04in Discord. We're going to make a video
1:13:05as the best practice. We'll make a video
1:13:06of the worst practice and the best
1:13:08practice. Worst practice is you're
1:13:09putting yourself in the queue and you
1:13:10sit quietly upset for 6 hours. That's
1:13:12the worst practice. Best practice, I'm
1:13:14in Discord. Hey folks, anybody else
1:13:16working on something? And you're
1:13:17chatting and you're talking and the TA
1:13:18might pull somebody off off and have a
1:13:19one-on-one thing and come back but
1:13:21you're chatting and you're talking and
1:13:22you're having a conversation. So, try to
1:13:23make Discord as productive as possible
1:13:25and as simulated of the face-to-face
1:13:28talking to each other in the hallway as
1:13:29we can. That's the whole idea and with
1:13:31that Hopefully, that video will help in
1:13:32terms of doing that. But thank you for
1:13:33that. And I I get it that there are
1:13:35things that is And one more thing to to
1:13:36emphasize.
1:13:37Yeah. We we are way over our TA hours.
1:13:40So, you know, we are burning a lot of TA
1:13:43hours so we can't throw in any more
1:13:45TA resources. We are trying to free them
1:13:47up as well. I mean, our second purpose
1:13:49of this reduce workload. We are trying
1:13:51to reduce a bit
1:13:52the number of of TA hours that we owe to
1:13:55to our TAs.
1:13:56Um but it it's there. We are trying to
1:13:59streamline this as much as we can.
1:14:03Uh let's see.
1:14:07Can I send you a time lapse workspace
1:14:09and a regular zoom video? All those
1:14:11things. If you want to send multiple
1:14:12things, that's great. At the very least,
1:14:13the bottom line is I got to be able to
1:14:15see your screen and see your desktop.
1:14:16So, it's not just well, here's the
1:14:17screen only. No, I need to see what your
1:14:19hands were doing. So, it's If you do a
1:14:20time lapse from the side, I can almost
1:14:22universally see screen and desktop and
1:14:24hands and phone and all those things.
1:14:26So, do what you need to do and if it
1:14:27means you need to submit multiple
1:14:28videos, that's a kind of answer before,
1:14:31please do that but it's not required if
1:14:32I can get to see everything there. So,
1:14:34that's the key. So,
1:14:35choose The more the better, to be
1:14:37honest. The more the better is what
1:14:38really we say there. Um
1:14:42Um
1:14:45I'm not sure I understand this one. Can
1:14:47we use the 15 minutes like a sort of 15
1:14:49minute mandatory
1:14:51office hour for us
1:14:54ask questions about project work
1:14:56especially.
1:14:5715 minutes per student is definitely a
1:14:59huge TA investment.
1:15:01And office hour queue has been long. I
1:15:04think it would be huge to use this like
1:15:06office hour to lower the straight
1:15:09strain on the office hour TAs. Yeah, you
1:15:12can use those 15 minutes whatever you
1:15:14like and related to
1:15:16Yeah. Yeah.
1:15:17Anyway you want to. Now, do you know Do
1:15:18know that we when we have project office
1:15:20hours, the TAs practice. They read the
1:15:22project. Not every TA knows every
1:15:24project. So, going to be very honest. Um
1:15:26all the staff is going to be supporting
1:15:27in this office hours but you may get a
1:15:29TA who doesn't know the project as well
1:15:31as you do. So, you know, they'll do
1:15:32their best and you'll teach them what it
1:15:33is but it might not be as effective as
1:15:35it would to project office hours. I'll
1:15:36just be very frank. Our TAs who
1:15:37run project office hours, they are rock
1:15:39star rock stars about that project.
1:15:40They've spent time reading it, doing it,
1:15:42answering questions about it. But a
1:15:44random TA doing random office hours may
1:15:46not know what you what you need. That's
1:15:47So, there's a little bit of like
1:15:49your mileage may vary.
1:15:51So, the question here, if you have
1:15:52personal questions, come to Dan and me.
1:15:55We have private
1:15:56uh
1:15:57slots every week.
1:15:58one-on-one time every week. Well, if we
1:16:00need more if we need more, you know,
1:16:02they have not been sold out yet.
1:16:04We can add more. So, we can talk to 12
1:16:07people per week.
1:16:08Amen. Amen on that. Uh
1:16:10more stuff. Um will mid midterm parts be
1:16:13labeled? Yes, they will. Which I I
1:16:15mentioned that to my students to my
1:16:16folks building this exam. We're going to
1:16:17label them very clearly that here's the
1:16:19topic. See this right here? It'll be
1:16:20like these are the five labels you're
1:16:21going to see for the five sections. The
1:16:24midterm sections. The word is sections,
1:16:26not questions. Sections. These sections
1:16:27will all be labeled. Great question. All
1:16:30right, good stuff.
1:16:32Um
1:16:34I had thought that we did the okayed to
1:16:36release the quest Bora, the quest
1:16:38solutions. Now,
1:16:39um
1:16:40Uh I think we are waiting for the for
1:16:42the the quest retake to be done. So, we
1:16:45are waiting for
1:16:45We were waiting for the quest retake cuz
1:16:47we don't do that. That was the idea and
1:16:48I think that now that that's done, I
1:16:50think we can really release that um and
1:16:52release some solutions. I'm not sure
1:16:53we'll release every test we use cuz we
1:16:55may be able to use discussions again but
1:16:56we're basically
1:16:58Here's the Here's a TA solution that
1:16:59would have gotten full credit. Think
1:17:00that'd be useful for you, right? I don't
1:17:01mind the TA solution. That'd be fine.
1:17:04Um
1:17:06So, this is a good question. Of the two
1:17:07me We need to have a conversation. How
1:17:09many slip days would you be using to fix
1:17:102B? Um I don't know how long it takes
1:17:12you to do that and how long we say to do
1:17:14this. So, it's a it's a little bit of We
1:17:16need to talk about the details of this
1:17:17one. Um
1:17:20Yeah, we already have apparently a Slack
1:17:22thread going on that.
1:17:25Yeah.
1:17:26So, we'll we'll get back to you. Yeah,
1:17:28we'll get back to I I still like it. So,
1:17:30what we need to talk about being being
1:17:31fair for other students.
1:17:32you know, we need to make make sure that
1:17:34it's fair. So, it you know, there'll be
1:17:35some some penalty in in slip days. We
1:17:38need to decide how many. Yeah, but
1:17:40that's a minor thing. If you saved all
1:17:41them up, you have a lot of them for
1:17:42that. That's great. Um
1:17:45Project office hours this week. Nothing
1:17:46on the calendar. That's interesting. I
1:17:48think this is most week was mostly going
1:17:49to be a setup for If I look at the the
1:17:52calendar, setup for the exam. And you're
1:17:54right. I don't see any project office
1:17:56hours here. I see a law section which is
1:17:57great.
1:17:58Because
1:17:59we are adding project
1:18:02midterm reviews. That's where we use
1:18:04those some of those hours.
1:18:05Yeah, so we burned We moved the to time
1:18:07over. Right. I I did I did see that.
1:18:09Yeah.
1:18:10Uh I I'm going to I'm going to say I
1:18:12don't know if we can We can maybe ask
1:18:14our TAs but in if we end up not adding
1:18:16any more, I would say go to all the
1:18:17office hours that you have to try to get
1:18:19support in the project space for that.
1:18:21Um and that's a little bit harder for
1:18:22that but I appreciate It doesn't doesn't
1:18:23answer all your questions but but I do
1:18:24that.
1:18:25Um
1:18:28Uh And if you use all your slip days
1:18:31because you were not feeling well, then
1:18:33you have no slip days and so that's
1:18:34Whenever you have no slip days, you
1:18:34can't do the slip which means you cannot
1:18:36uh slip on a project to add more time
1:18:38for that and you can't slip on doing the
1:18:40redo of 2B. So, that's that's base of
1:18:42that.
1:18:42We We have added just for reference, if
1:18:44you were sick and I have a doctor's
1:18:47note, send it to us. That gets you We
1:18:50can give you extras. If If you actually
1:18:51formally sick that like I went to the
1:18:53hospital, I was here today and I was
1:18:54out, then like we can actually do that
1:18:56uh and Bora, I don't know if you want to
1:18:57manage that via form or via email to
1:18:59you. I certainly don't want any more
1:19:00email cuz I I can't do it.
1:19:01We did We did it by a form. We do it by
1:19:03a form.
1:19:05So, the form for that. I don't want to
1:19:07I can't do it as it is.
1:19:08Um
1:19:10Uh
1:19:12shubba dubba This is about the spec
1:19:13being clear.
1:19:15Um
1:19:17I think we as a question about the spec
1:19:19that try to update the spec Well, not
1:19:22the spec but try to up- update the auto
1:19:24grader because the spec wasn't clear and
1:19:25I think that already happened. I mean we
1:19:27can bring that particular question back
1:19:29to the TAs. I wasn't in the lower level
1:19:31loop of that conversation. I just said
1:19:33hey guys fix this. You shouldn't be able
1:19:34to lose 40 points on a small error and
1:19:37so I think that already happened but
1:19:39let's let's make sure let me I'm just
1:19:41copy and paste this into my channel just
1:19:43to make sure that this one is resolved
1:19:44because I don't have the answer for that
1:19:45one right easy here. Okay, I'll say a
1:19:48student asked
1:19:52and I will try to get back to you on the
1:19:54on that on that last one next week.
1:19:55We're here to answer these questions.
1:19:57We'll try to be here again next week. I
1:19:58actually don't mind staying and
1:19:59answering all the questions that are
1:19:59there. It's like our face-to-face time
1:20:01with everybody. This is great. For those
1:20:03who decide to come to the section which
1:20:04we thank you the folks you know how many
1:20:05people came so and and that live. Um
1:20:09Uh
1:20:11Yeah, I think that's the same kind of
1:20:12thing. If the auto grader chance that we
1:20:13fail one of the tests in a group do you
1:20:15all the test group taken off? I don't
1:20:16think that's true. I think I think the
1:20:17way that it's supposed to work is
1:20:19that that that used to be that used to
1:20:21be like that but we have changed that.
1:20:22So Steven managed to break this up.
1:20:25Perfect. Perfect. And thanks to Steven
1:20:26for the hard work to do that. Yeah, it
1:20:28was a little bit rough of a of a of a
1:20:30cut but seems to be working. Yeah,
1:20:33perfect. All right folks, thank you so
1:20:34much for showing up. 5:30 we're 30
1:20:36minutes over. We're here for you, okay?
1:20:38Hopefully those that that that relief
1:20:40package helps a lot. We'll see you guys
1:20:41next week. Thank you folks.
1:20:44All right. Any questions? Thanks folks.
1:20:46Okay, bye.

About this transcript

This page contains the full transcript of [CS61C FA20] Weekly Lecture 09.LIVE - Cache by CS 61C Departmental, generated from the public captions YouTube serves with the video. The transcript has 17,397 words across 2,575 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.