kvcache-ai/ktransformers

Fork 0

mirror of https://github.com/kvcache-ai/ktransformers.git synced 2026-11-05 01:07:28 +00:00

[GH-ISSUE #521] 内存占用上不去 #374

New issue

Closed

opened 2026-10-29 14:38:45 +00:00 by bsakaguchi · 5 comments

bsakaguchi commented

2026-10-29 14:38:45 +00:00

Owner

Originally created by @zzk2021 on GitHub (Feb 20, 2025).
Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/521

top - 10:42:11 up 21:34, 1 user, load average: 28.63, 20.14, 13.18
Tasks: 1384 total, 4 running, 1375 sleeping, 4 stopped, 1 zombie
%Cpu(s): 8.1 us, 10.0 sy, 0.0 ni, 81.3 id, 0.3 wa, 0.2 hi, 0.0 si, 0.0 st
MiB Mem : 515355.1 total, 238953.3 free, 6869.1 used, 269532.6 buff/cache
MiB Swap: 4096.0 total, 974.9 free, 3121.1 used. 505078.5 avail Mem

速度只有1.3token/s, 16x32 ddr4 3200mhz，志强3代，kt版本0.21
跑的是q4量化版本

Originally created by @zzk2021 on GitHub (Feb 20, 2025). Original GitHub issue: https://github.com/kvcache-ai/ktransformers/issues/521 top - 10:42:11 up 21:34, 1 user, load average: 28.63, 20.14, 13.18 Tasks: 1384 total, 4 running, 1375 sleeping, 4 stopped, 1 zombie %Cpu(s): 8.1 us, 10.0 sy, 0.0 ni, 81.3 id, 0.3 wa, 0.2 hi, 0.0 si, 0.0 st MiB Mem : 515355.1 total, 238953.3 free, 6869.1 used, 269532.6 buff/cache MiB Swap: 4096.0 total, 974.9 free, 3121.1 used. 505078.5 avail Mem 速度只有1.3token/s, 16x32 ddr4 3200mhz，志强3代，kt版本0.21 跑的是q4量化版本

bsakaguchi

2026-10-29 14:38:45 +00:00

closed this issue
added the
stale
label

bsakaguchi commented

2026-10-29 14:38:46 +00:00

Author

Owner

@zzk2021 commented on GitHub (Feb 20, 2025):

@zzk2021 commented on GitHub (Feb 20, 2025): ![Image](https://github.com/user-attachments/assets/cf4a1ef3-eb25-4f09-b31b-88eb0e98ced7)

bsakaguchi commented

2026-10-29 14:38:46 +00:00

Author

Owner

@yileld commented on GitHub (Feb 20, 2025):

请问你的 mbw 100 命令显示的内存带宽是多少？

@yileld commented on GitHub (Feb 20, 2025): 请问你的 mbw 100 命令显示的内存带宽是多少？

bsakaguchi commented

2026-10-29 14:38:46 +00:00

Author

Owner

@zzk2021 commented on GitHub (Feb 20, 2025):

请问你的 mbw 100 命令显示的内存带宽是多少？

我用stream测试有300多Gib，但是用mbw最高只有10000多Mib，差别很大，不知道为什么

@zzk2021 commented on GitHub (Feb 20, 2025): > 请问你的 mbw 100 命令显示的内存带宽是多少？我用stream测试有300多Gib，但是用mbw最高只有10000多Mib，差别很大，不知道为什么

bsakaguchi commented

2026-10-29 14:38:46 +00:00

Author

Owner

@yileld commented on GitHub (Feb 20, 2025):

请问你的 mbw 100 命令显示的内存带宽是多少？

我用stream测试有300多Gib，但是用mbw最高只有10000多Mib，差别很大，不知道为什么

搜了下stream是理论带宽，所以会差这么多。官方用的理论带宽好像是600G

@yileld commented on GitHub (Feb 20, 2025): > > 请问你的 mbw 100 命令显示的内存带宽是多少？ > > 我用stream测试有300多Gib，但是用mbw最高只有10000多Mib，差别很大，不知道为什么搜了下stream是理论带宽，所以会差这么多。官方用的理论带宽好像是600G

bsakaguchi commented

2026-10-29 14:38:46 +00:00

Author

Owner

@ErvinXie commented on GitHub (Dec 11, 2025):

This issue has been automatically closed due to inactivity for 60 days.

If you believe this issue is still relevant, please feel free to reopen it with additional information or context.

Thank you for your contribution to KTransformers!

此 issue 因超过 60 天无活动而被自动关闭。

如果您认为此问题仍然存在，请随时重新打开并提供更多信息或上下文。

感谢您对 KTransformers 的贡献！

@ErvinXie commented on GitHub (Dec 11, 2025): This issue has been automatically closed due to inactivity for 60 days. If you believe this issue is still relevant, please feel free to reopen it with additional information or context. Thank you for your contribution to KTransformers! --- 此 issue 因超过 60 天无活动而被自动关闭。如果您认为此问题仍然存在，请随时重新打开并提供更多信息或上下文。感谢您对 KTransformers 的贡献！

bsakaguchi referenced this issue

2026-10-29 14:40:47 +00:00

[GH-ISSUE #712] Does the official plan support Turing Nvidia GPU ？ #521

bsakaguchi referenced this issue

2026-10-29 14:41:28 +00:00

[GH-ISSUE #801] [Bug] RuntimeError: CUDA error: invalid device function #580

bsakaguchi referenced this issue

2026-10-29 14:44:57 +00:00

[GH-ISSUE #1192] [Bug] 在Ubuntu下编译失败 #818