mrfelek a293f7901b perf(agc): stage written-only swept elements on discrete GPUs ...

On a discrete GPU a write-only V# element whose stride is wider than the
widest buffer store (16 bytes) and that a direct dispatch sweeps is staged
in a device-local shadow of up to 256 MiB, as read and written swept
elements are. Its stores no longer cross the bus into the host import.

AI-assisted. Tested by me on my own machine.

2026-10-24 10:22:12 +00:00

..

debt

perf(agc): stage written-only swept elements on discrete GPUs

2026-10-24 10:22:12 +00:00

ARCHITECTURE.md

Merge pull request #8527 from eliasbruno124-dev/fix/host-import-settle-order-20261020

2026-10-23 20:59:54 +02:00

BUILD.md

Merge pull request #9016 from dantdmdarealguy/docs/build-python-launcher-windows

2026-10-24 11:35:41 +02:00

CONVENTIONS.md

docs(dev): require a HostWrite bracket around host writes into guest memory

2026-10-21 14:57:32 +13:00

ENVIRONMENT.md

Merge pull request #7307 from Zenodeon/agc-submit-progress-pr

2026-10-24 11:59:57 +02:00

GAME_TESTING.md

Merge pull request #8807 from ycbuzi/fix/agc-driver-frame-dump-wait

2026-10-22 17:50:26 +13:00

HW_ORACLE.md

feat(hw-oracle): run on Windows through the AMD driver's HIP runtime and comgr

2026-10-21 11:07:06 -04:00

PROGRESS.md

…

REVERSE_ENGINEERING.md

docs(dev): guide to reading call sites and library disassembly

2026-10-21 11:56:11 +02:00

TechnicalDebt.md

Merge pull request #8241 from Henryisdjfunkyho/fix/check-compile-response-files

2026-10-21 15:02:36 +13:00

TRACING.md

…