igneum/proto-cuda/packs-readwidth/scr4k128
2026-10-05 20:18:35 +00:00
..
kernel.cl read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
kernel.cu read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
kernel_bound.cl read-width: OpenCL scratch kernels declare the __local exchange buffer before the unit loop (AMD's compiler requires the outermost scope; round 3 on the 9070 XT); seven scratch packs re-exported, vectors unchanged 2026-10-05 20:18:35 +00:00
kernel_bound.cu read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
memhard.h read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
memhard.metal read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
program.h read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
program.json read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
program.metal read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
program_bound.metal read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
vectors.h read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00
vectors.json read-width: scratch per warp is a class parameter (32 or 128 KiB, under the 6 GB working-set cap), distinct-address rule bounds dataset loads only; Metal pack harness; OpenCL --bench-pack, scratch args and 16-byte probe; packfile class fields; OpenCL emulator persistent launch 2026-10-05 19:58:20 +00:00