Skip to content

[None][doc] GVR V2: Self-Sampling and Multi-Thresholding for Faster Exact Top-K #19244

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
longcheng-nv merged 33 commits into from
Sep 22, 2026
Merged

[None][doc] GVR V2: Self-Sampling and Multi-Thresholding for Faster Exact Top-K #19244

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
33 commits
Select commit Hold shift + click to select a range
fcbb1ea
[None][doc] Explain GVR V2 self-sampling and multi-threshold Top-K
longcheng-nv Sep 16, 2026
329cc14
[None][doc] Add DeepSelect FP32 speedup heatmap to GVR V2 blog
longcheng-nv Sep 16, 2026
c747dc3
[None][doc] Keep GVR V2 blog focused on FP32 comparisons
longcheng-nv Sep 16, 2026
02fbcf8
[None][doc] Compare average and peak reach of GVR V2 Pareto curves
longcheng-nv Sep 16, 2026
73620c2
[None][doc] Explain shared GVR V2 kernel integration for decode and p…
longcheng-nv Sep 16, 2026
a891ef1
[None][doc] Clarify GVR V2 narrative with recovery and integration di…
longcheng-nv Sep 16, 2026
ab08761
[None][doc] Adopt GVR V2 algorithm title and unified-core subtitle
longcheng-nv Sep 17, 2026
8560db0
[None][doc] Explain robustness and integration motives for GVR V2
longcheng-nv Sep 17, 2026
773914b
[None][doc] Connect GVR V2 to SELECT and visualize temporal overlap
longcheng-nv Sep 17, 2026
1491efd
[None][doc] Cover GVR V2 dispatch updates and widen overlap figure
longcheng-nv Sep 17, 2026
50be207
[None][doc] Quantify temporal hint variability across inputs and layers
longcheng-nv Sep 17, 2026
9ad0776
[None][doc] Organize the GVR V2 blog into four main sections
longcheng-nv Sep 17, 2026
bd66cd6
[None][doc] Refresh GVR V2 benchmarks with PR 19076 measurements
longcheng-nv Sep 17, 2026
4b0f7d4
[None][doc] Explain GVR V2 threshold quality and candidate work
longcheng-nv Sep 17, 2026
452ce69
[None][doc] Keep candidate figure math outside caption emphasis
longcheng-nv Sep 17, 2026
9a72d53
[None][doc] Detail GPU self-sampling windows and calibration
longcheng-nv Sep 17, 2026
0788779
[None][doc] Correct later temporal V1 multi-threshold comparison
longcheng-nv Sep 17, 2026
3ee254c
[None][doc] Align GVR blog claims with main and measurement scope
longcheng-nv Sep 17, 2026
491b95b
[None][doc] Clarify threshold and candidate work in Figure 4
longcheng-nv Sep 17, 2026
17d920f
[None][doc] Use trtllm-bench in GVR V2 enablement example
longcheng-nv Sep 17, 2026
d2d31f0
[None][doc] Align heatmap introduction with plotted baselines
longcheng-nv Sep 17, 2026
3a6f90b
[None][doc] Strengthen GVR V2 narrative and performance evidence
longcheng-nv Sep 17, 2026
66324a4
[None][doc] Focus GVR V2 comparisons on four baselines
longcheng-nv Sep 17, 2026
47b79f8
[None][doc] Focus GVR V2 performance on V1 and native radix
longcheng-nv Sep 18, 2026
1b363f0
[None][doc] Keep a single temporal GVR V1 comparison
longcheng-nv Sep 18, 2026
2473b5a
[None][doc] Clarify temporal-hint index mapping in Figure 2
longcheng-nv Sep 18, 2026
7ed8feb
[None][doc] Simplify Figure 5 candidate population diagram
longcheng-nv Sep 18, 2026
3d40422
[None][doc] Explain GVR V1 local wins and center blog tables
longcheng-nv Sep 18, 2026
6642d61
[None][doc] Add GVR V1 speedup heatmap to GVR V2 blog
longcheng-nv Sep 19, 2026
4df06da
[None][docs] Address timing parsing and clarify GVR prior contracts
longcheng-nv Sep 19, 2026
b157334
[None][doc] Treat JSON blog data as downloadable assets
longcheng-nv Sep 21, 2026
c738302
docs: use TensorRT LLM branding in GVR V2 blog
longcheng-nv Sep 22, 2026
b829727
docs: link GVR V2 references to the official documentation
longcheng-nv Sep 22, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
188 changes: 188 additions & 0 deletions docs/source/blogs/media/gvr_v2/README.md
View file
Open in desktop

Large diffs are not rendered by default.

762 changes: 762 additions & 0 deletions docs/source/blogs/media/gvr_v2/algorithm.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
690 changes: 690 additions & 0 deletions docs/source/blogs/media/gvr_v2/candidate_work.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
374 changes: 374 additions & 0 deletions docs/source/blogs/media/gvr_v2/evolution.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/blogs/media/gvr_v2/flash_timings.csv.gz
View file
Open in desktop
Binary file not shown.
469 changes: 469 additions & 0 deletions docs/source/blogs/media/gvr_v2/gpu_sampling.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
1,770 changes: 1,770 additions & 0 deletions docs/source/blogs/media/gvr_v2/gvr_v1_map.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
292 changes: 292 additions & 0 deletions docs/source/blogs/media/gvr_v2/integration.svg
View file
Open in desktop
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Loading