Skip to content
Navigation Menu
Sign in
Appearance settings
Platform
AI CODE CREATION
GitHub Copilot
Write better code with AI
GitHub Copilot app
Direct agents from issue to merge
MCP Registry
Integrate external tools
DEVELOPER WORKFLOWS
Actions
Automate any workflow
Codespaces
Instant dev environments
Issues
Plan and track work
Code Review
Manage code changes
Code Quality
Enforce quality at merge
APPLICATION SECURITY
GitHub Advanced Security
Find and fix vulnerabilities
Code security
Secure your code as you build
Secret protection
Stop leaks before they start
EXPLORE
Why GitHub
Documentation
Blog
Changelog
Marketplace
View all features
Solutions
BY COMPANY SIZE
Enterprises
Small and medium teams
Startups
Nonprofits
BY USE CASE
App Modernization
DevSecOps
DevOps
CI/CD
View all use cases
BY INDUSTRY
Healthcare
Financial services
Manufacturing
Government
View all industries
View all solutions
Resources
EXPLORE BY TOPIC
AI
Software Development
DevOps
Security
View all topics
EXPLORE BY TYPE
Customer stories
Events & webinars
Ebooks & reports
Business insights
GitHub Skills
SUPPORT & SERVICES
Documentation
Customer support
Community forum
Trust center
Partners
View all resources
Open Source
COMMUNITY
GitHub Sponsors
Fund open source developers
PROGRAMS
Security Lab
Maintainer Community
GitHub Stars
Archive Program
REPOSITORIES
Topics
Trending
Collections
Enterprise
ENTERPRISE SOLUTIONS
Enterprise platform
AI-powered developer platform
AVAILABLE ADD-ONS
GitHub Advanced Security
Enterprise-grade security features
Copilot for Business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
C2SM
/
icon4py
Public
Notifications
You must be signed in to change notification settings
Fork
12
Star
21
Code
Issues
66
Pull requests
54
Actions
Projects
Security and quality
0
Insights
Additional navigation options
Code
Issues
Pull requests
Actions
Projects
Security and quality
Insights
Optimize
compute_zdiff_gradp
- #1272
#1272
Draft
msimberg
wants to merge 46 commits into
C2SM:main
C2SM/icon4py:main
from
msimberg:zdiff-vertoffset-gradp
msimberg/icon4py:zdiff-vertoffset-gradp
Copy head branch name to clipboard
Conversation
Commits
46
(46)
Checks
Files changed
Draft
Optimize
compute_zdiff_gradp
#1272
msimberg
wants to merge 46 commits into
C2SM:main
C2SM/icon4py:main
from
msimberg:zdiff-vertoffset-gradp
msimberg/icon4py:zdiff-vertoffset-gradp
Copy head branch name to clipboard
Commits
Commits on May 19, 2026
perf(metrics): replace nested python loops with np.searchsorted in compute_zdiff_gradp
Show description for d725195
msimberg
committed
d725195
View commit details
Copy full SHA for d725195
Browse repository at this point
perf(metrics): eliminate vertidx_gradp intermediate array in compute_zdiff_gradp
Show description for 9aade50
msimberg
committed
9aade50
View commit details
Copy full SHA for 9aade50
Browse repository at this point
perf(metrics): pre-compute ascending z_ifc and unroll per-side loops
Show description for bc8f00c
msimberg
committed
bc8f00c
View commit details
Copy full SHA for bc8f00c
Browse repository at this point
test(metrics): remove cpu_only mark from test_factory_zdiff_gradp
Show description for a496ec1
msimberg
committed
a496ec1
View commit details
Copy full SHA for a496ec1
Browse repository at this point
perf(metrics): batched 2D searchsorted via row-offset trick (Divakar)
Show description for 411a961
msimberg
committed
411a961
View commit details
Copy full SHA for 411a961
Browse repository at this point
refactor(metrics): use array_ns instead of manual backend detection
Show description for 86e95db
msimberg
committed
86e95db
View commit details
Copy full SHA for 86e95db
Browse repository at this point
perf(metrics): remove z_me mask, use do_exchange for zdiff_gradp halo
Show description for 0bff01e
msimberg
committed
0bff01e
View commit details
Copy full SHA for 0bff01e
Browse repository at this point
perf(metrics): remove internal z_me exchange, drop exchange parameter
Show description for 1706d16
msimberg
committed
1706d16
View commit details
Copy full SHA for 1706d16
Browse repository at this point
Only run common mpi tests
msimberg
committed
7afb739
View commit details
Copy full SHA for 7afb739
Browse repository at this point
Commits on May 20, 2026
Refactor compute_zdiff_gradp to only do do_exchange exchange
msimberg
committed
84bb917
View commit details
Copy full SHA for 84bb917
Browse repository at this point
Some more refactoring
msimberg
committed
ffa6397
View commit details
Copy full SHA for ffa6397
Browse repository at this point
Merge remote-tracking branch 'origin/main' into zdiff-vertoffset-gradp
msimberg
committed
1d48eeb
View commit details
Copy full SHA for 1d48eeb
Browse repository at this point
Commits on Jun 1, 2026
Apply pre-commit auto-fixes from merge
msimberg
committed
6ca86c9
View commit details
Copy full SHA for 6ca86c9
Browse repository at this point
Commits on Aug 24, 2026
Add compute_zdiff_gradp_v2 with float64 searchsorted, gated validation, and endpoint fallback
Show description for 2e5c5a7
msimberg
committed
2e5c5a7
View commit details
Copy full SHA for 2e5c5a7
Browse repository at this point
[skip ci] iter-002: exact variant, always-on validation, tests
msimberg
committed
e0c9069
View commit details
Copy full SHA for e0c9069
Browse repository at this point
[skip ci] fix: on-device v2 reductions, exact int32 fallback removal, round-robin benchmark, NaN test fix, hoist invariant [R30-R34]
msimberg
committed
456bef4
View commit details
Copy full SHA for 456bef4
Browse repository at this point
[skip ci] untrack .build-skill/benchmark (build-skill working artifact)
msimberg
committed
27b05ad
View commit details
Copy full SHA for 27b05ad
Browse repository at this point
[skip ci] fix: replace minimum.accumulate with Hillis-Steele doubling scan for cupy compatibility
msimberg
committed
2e59333
View commit details
Copy full SHA for 2e59333
Browse repository at this point
[skip ci] iter-003 (D1): compute_zdiff_gradp_exact_v2 with zero assumed premises
msimberg
committed
b32e945
View commit details
Copy full SHA for b32e945
Browse repository at this point
[skip ci] iter-004 (B1): compute_zdiff_gradp_dispatch and shared v2 validation
msimberg
committed
684b866
View commit details
Copy full SHA for 684b866
Browse repository at this point
[skip ci] fix: memory-capped auto-chunking for compute_zdiff_gradp_exact_v2
msimberg
committed
8c75b03
View commit details
Copy full SHA for 8c75b03
Browse repository at this point
Commits on Aug 25, 2026
[skip ci] compute_zdiff_gradp_exact_v3 product code and unit tests
Show description for d6ed9d1
msimberg
committed
d6ed9d1
View commit details
Copy full SHA for d6ed9d1
Browse repository at this point
[skip ci] benchmark wiring for compute_zdiff_gradp_exact_v3
Show description for ccbd5be
msimberg
committed
ccbd5be
View commit details
Copy full SHA for ccbd5be
Browse repository at this point
[skip ci] untrack .build-skill/benchmark (build-skill working artifact)
msimberg
committed
93e43c2
View commit details
Copy full SHA for 93e43c2
Browse repository at this point
fix: gate exact_v2/exact_v3 validation; stack v2 validation [R49-R52] [skip ci]
msimberg
committed
4117f6a
View commit details
Copy full SHA for 4117f6a
Browse repository at this point
feat: detect exact interior ties in compute_zdiff_gradp v2 validation and route dispatch to exact fallback [skip ci]
msimberg
committed
792dcb0
View commit details
Copy full SHA for 792dcb0
Browse repository at this point
feat: premise-free GT4Py compute_zdiff_gradp variant (embedded; gtfn codegen blocked) [skip ci]
msimberg
committed
169b7ed
View commit details
Copy full SHA for 169b7ed
Browse repository at this point
fix: exact GT4Py carry via per-level first-match re-invocation [R53-R57] [skip ci]
msimberg
committed
9dacbd4
View commit details
Copy full SHA for 9dacbd4
Browse repository at this point
fix: eliminate tie-check double work by reusing fast-path searchsorted positions [skip ci]
msimberg
committed
57dccd4
View commit details
Copy full SHA for 57dccd4
Browse repository at this point
fix: R59 nlev-1 endpoint tie detection, R60/R61 cleanup, F4 regression test [skip ci]
msimberg
committed
14c93eb
View commit details
Copy full SHA for 14c93eb
Browse repository at this point
Add compute_zdiff_gradp_exact_v4: premise-free broadcast first-match [skip ci]
msimberg
committed
ea230f6
View commit details
Copy full SHA for ea230f6
Browse repository at this point
Commits on Aug 26, 2026
fix: address v4 review findings R62-R64, F5, and F6 comment [R62][R63][R64][F5][F6] [skip ci]
msimberg
committed
a71b3b2
View commit details
Copy full SHA for a71b3b2
Browse repository at this point
feat: add compute_zdiff_gradp_exact_v5 with cupy ElementwiseKernel fast path [skip ci]
msimberg
committed
d40467c
View commit details
Copy full SHA for d40467c
Browse repository at this point
test: extend exact_v5 coverage to all parametrized lists and add kernel compile test [skip ci]
msimberg
committed
881dd71
View commit details
Copy full SHA for 881dd71
Browse repository at this point
Commits on Aug 27, 2026
[skip ci] Replace compute_zdiff_gradp with premise-free literal main implementation
Show description for e58bacf
msimberg
committed
e58bacf
View commit details
Copy full SHA for e58bacf
Browse repository at this point
[skip ci] feat: add compute_zdiff_gradp_fast parallel successor-table variant
Show description for 457a988
msimberg
committed
457a988
View commit details
Copy full SHA for 457a988
Browse repository at this point
[skip ci] perf: precompute constant-query bracket in _carry_first_match
Show description for 9b1509f
msimberg
committed
9b1509f
View commit details
Copy full SHA for 9b1509f
Browse repository at this point
[skip ci] refactor compute_zdiff_gradp to single searchsorted-based function
Show description for 60042ef
msimberg
committed
60042ef
View commit details
Copy full SHA for 60042ef
Browse repository at this point
[skip ci] update compute_zdiff_gradp tests for searchsorted implementation
Show description for 2e7f2af
msimberg
committed
2e7f2af
View commit details
Copy full SHA for 2e7f2af
Browse repository at this point
[skip ci] remove unused fill_low
msimberg
committed
6b3d87c
View commit details
Copy full SHA for 6b3d87c
Browse repository at this point
[skip ci] fix: restore side=left for GPU, add uniform tie correction
Show description for 499647f
msimberg
committed
499647f
View commit details
Copy full SHA for 499647f
Browse repository at this point
Commits on Aug 28, 2026
[skip ci] simplify: remove tie correction, document premise
Show description for ee9861c
msimberg
committed
ee9861c
View commit details
Copy full SHA for ee9861c
Browse repository at this point
Commits on Aug 29, 2026
Merge origin/main into zdiff-vertoffset-gradp
msimberg
committed
b2d6706
View commit details
Copy full SHA for b2d6706
Browse repository at this point
fix: remove unused decomposition import after merge [skip ci]
msimberg
committed
55be2af
View commit details
Copy full SHA for 55be2af
Browse repository at this point
fix: remove dead CandDim (R65) and avoid GPU host syncs in fill/scalar reduction (R66)
msimberg
committed
93a3264
View commit details
Copy full SHA for 93a3264
Browse repository at this point
Commits on Aug 31, 2026
docs: replace figurative "carry" with "lower-bound update" and split long sentences (R71, R72) [skip ci]
msimberg
committed
2d0d538
View commit details
Copy full SHA for 2d0d538
Browse repository at this point
You can’t perform that action at this time.