[GH-ISSUE #1332] Task: Labs checks #4375

New Issue

GiteaMirror · 2026-04-19T12:23:48-05:00

GiteaMirror commented

2026-04-19 12:23:48 -05:00

Originally created by @asgalon on GitHub (Apr 14, 2026).
Original GitHub issue: https://github.com/harvard-edge/cs249r_book/issues/1332

To make the handling a bit easier, this ticket collects documentation about bugs / improvements / todos with the marimo labs from vol1/lab02 on.

Repeating issues to check:

Are the prediction choice menues displayed to make it possible to make a prediction and unlock the lower parts?

Vol. 1

Lab 02

missing prediction choices
String formatting issue, "Upgrading from A100 ($15K) to H100 ($30K)" has to be written "Upgrading from A100 (\\$15K) to H100 (\\$30K)" in the label or else it will be rendered as Upgrading from A100 (15K) to H100 (30K) as partly italic
Part D structure access/modification error, move partD_data_size and partD_wireless one cell up
Part A tab: The text in the yellow box should adapt the AI value from the slider, or keep the speedup value at the 5 FLOPS/Byte point. Right now, the speedup value changes with the slider, but the 5 is fixed.

Latency improvement value / speedup calculations are a bit mismatched. Selecting 1.1x above leads to a correct 1.64 result below. This is a bit confusing, at least to me... Updated example values, perhaps?

Lab 03

missing prediction choices
Part C Allocate your Team's 10 PMs - Chart stuck, does not move when changing items

Originally created by @asgalon on GitHub (Apr 14, 2026). Original GitHub issue: https://github.com/harvard-edge/cs249r_book/issues/1332 To make the handling a bit easier, this ticket collects documentation about bugs / improvements / todos with the marimo labs from vol1/lab02 on. Repeating issues to check: - Are the prediction choice menues displayed to make it possible to make a prediction and unlock the lower parts? ## Vol. 1 ### Lab 02 - missing prediction choices - String formatting issue, "Upgrading from A100 ($15K) to H100 ($30K)" has to be written `"Upgrading from A100 (\\$15K) to H100 (\\$30K)"` in the label or else it will be rendered as Upgrading from A100 (15K) to H100 (30K) as partly italic - Part D structure access/modification error, move partD_data_size and partD_wireless one cell up - Part A tab: The text in the yellow box should adapt the AI value from the slider, or keep the speedup value at the 5 FLOPS/Byte point. Right now, the speedup value changes with the slider, but the 5 is fixed. <img width="1795" height="983" alt="Image" src="https://github.com/user-attachments/assets/b3798516-8b02-4c57-acc5-aa5b9ab90602" /> - Latency improvement value / speedup calculations are a bit mismatched. Selecting 1.1x above leads to a correct 1.64 result below. This is a bit confusing, at least to me... Updated example values, perhaps? ### Lab 03 - missing prediction choices - Part C Allocate your Team's 10 PMs - Chart stuck, does not move when changing items

GiteaMirror added the area: book type: bug labels 2026-04-19 12:23:48 -05:00

GiteaMirror closed this issue

2026-04-19 12:23:50 -05:00

GiteaMirror commented

2026-04-19 12:23:52 -05:00

@profvjreddi commented on GitHub (Apr 17, 2026):

all concrete items from this thread have landed across four PRs — closing out.

PR	scope
#1385	lab_02 + lab_03 widget dataflow, $15K escape, AI=5 anchor, part D cell move
#1386	all-labs (33) widget return-tuple sweep + static regression guard
#1387	lab_02 Part A options/scenario aligned to real Hardware specs (3x not 6x; correct answer 1.6x not <1.1x); 42 tabs-cell widgets extracted to their own cells across 16 labs
#1374	browser-level WASM smoke (Playwright + COEP/COOP) as a separate regression guard against the underlying class of "widget invisible but no exception" bugs

static audit: 17 offending labs / 175 unreturned widgets → 0 / 0.

any further lab issues as they come up — please open fresh tickets so they don't get lost in a long thread.

thanks @asgalon for the careful debugging and screenshots on lab_00, lab_02, lab_03.

@profvjreddi commented on GitHub (Apr 17, 2026): all concrete items from this thread have landed across four PRs — closing out. | PR | scope | |-|-| | #1385 | lab_02 + lab_03 widget dataflow, \$15K escape, AI=5 anchor, part D cell move | | #1386 | all-labs (33) widget return-tuple sweep + static regression guard | | #1387 | lab_02 Part A options/scenario aligned to real Hardware specs (3x not 6x; correct answer 1.6x not <1.1x); 42 tabs-cell widgets extracted to their own cells across 16 labs | | #1374 | browser-level WASM smoke (Playwright + COEP/COOP) as a separate regression guard against the underlying class of "widget invisible but no exception" bugs | static audit: 17 offending labs / 175 unreturned widgets → **0 / 0**. any further lab issues as they come up — please open fresh tickets so they don't get lost in a long thread. thanks @asgalon for the careful debugging and screenshots on lab_00, lab_02, lab_03.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/cs249r_book#4375