mirror of
https://github.com/kvcache-ai/ktransformers.git
synced 2026-11-05 01:07:28 +00:00
[GH-ISSUE #1496] 模型大小140G,设备内存256G,不开启 NUMA,内存占用为什么只有4G左右? #1002
Labels
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
kvcache-ai/ktransformers#1002
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @corengh on GitHub (Sep 8, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/1496
CPU:INTEL XEON PLATINUM 8582C 2.6GHZ 60核
内存:DDR5-5600 32G × 8
显卡:NVIDIA GeForce RTX 3090 24G × 2
系统:Ubuntu 24.04
KTranformers 版本:0.3.2
加载模型:DeepSeek-R1-UD-IQ1_S(140.23G)
双路主板,但是目前只使用了一路 CPU。
编译时不使用
export USE_NUMA=1,运行时内存占用只有 4G 左右,显存占用 12-13G(不使用 balance)或 15-16G(使用 balance),decode 13 Tokens/s 左右。编译时开启 NUMA,运行时内存占用 236G 左右,其它方面包括性能表现和不开启没有差异,反而因为剩余内存较低,感觉有些卡顿。
求解的疑惑是:
@mrgaolei commented on GitHub (Sep 9, 2025):
看buffer/cache,主要内存占用在这里
@corengh commented on GitHub (Sep 9, 2025):
嗯,我也有观察到这块的变化,也就是加载到 buffer/cache 里是正常的是吗?但是为什么开启 NUMA 后,buffer/cache 没有增加,而占用都在 used 里?
@wqshmzh commented on GitHub (Sep 18, 2025):
要用free -h看cache的占用空间