0
Fork 0
mirror of https://github.com/zml/zml.git synced 2026-11-16 19:11:09 +00:00
zml/stdx
Guillaume Wenzek 1480f77a55
zml: add the platform tag inside the Buffer (#326)
CPU buffers behave differently from Cuda/Rocm/TPU buffers,
and in particular they are addressable from main memory.
This was annoying and prevented us to add some safety checks that
weren't valid for CPU.

I'm adding a tag to buffers so that we know which target they refer to.

Also there is a bug in PJRT unitialized which doesn't respected the
memory. I've opened https://github.com/openxla/xla/pull/31292
but in the meantime I've a workaround in ZML that uses the regular
.fromHostBuffer.
(I've also a patched version of PJRT plugins in
https://github.com/zml/pjrt-artifacts/pull/41 but not merged yet)
History 2025-09-18 15:35:25 +02:00
..
bounded_array.zig workspace: zig 0.15 (#316) 2025-08-27 14:11:22 +02:00
BUILD.bazel ci: revamp commit workflow (#320) 2025-09-03 14:23:22 +02:00
debug.zig zml: add stdx, rework async signature inference (#89) 2024-11-15 17:12:07 +01:00
flags.zig zml: add the platform tag inside the Buffer (#326) 2025-09-18 15:35:25 +02:00
fmt.zig zml: remove old writer interface APIs (#324) 2025-09-18 11:36:23 +02:00
fs.zig workspace: switch from zig_cc_binary to zig_binary (#311) 2025-08-21 13:12:15 +02:00
json.zig zml: add some buffer / hostbuffer helpers (#232) 2025-05-20 15:08:35 +02:00
math.zig workspace: add debug/release configurations and various stdx (#234) 2025-05-19 21:26:28 +02:00
meta.zig workspace: bump to zig=0.14.1 (#242) 2025-05-30 16:51:55 +02:00
queue.zig Bump rules_zig, add zig_srcs targets, cleanup build (#267) 2025-07-15 14:18:15 +02:00
signature.zig workspace: zig 0.15 (#316) 2025-08-27 14:11:22 +02:00
stdx.zig zml: loadById (#322) 2025-09-11 10:28:16 +02:00
time.zig workspace: zig 0.15 (#316) 2025-08-27 14:11:22 +02:00