[GPU] Optimize RMS Stack Size for Better Performance

[zaixing-wang]

Details:

• Since stack_size can be calculated, we can use smaller stack size instead of 33 to achieve better performance.

Performance:

• qwen2-0.5B’s throughput increased from 40.2 qps to 41.8 qps.

Tickets:

• CVS-152013

[dnkurek]

Hint: maybe check if dispatchData.dataSize is actually constant and won't change when executing LLM, therefore we know beforehand what will always dispatchData.itemsNum be in each execution. In other words, maybe the only thing that changes is dataCount

[zaixing-wang]

In Qwen, dataSize is always equal to 896.

[dnkurek]

I see, so that looks to me like a possible optimization there, since now you exactly know how much private memory you will need and better use the GPU's resources

[dnkurek]

And also you would know LWS and therefore you could use reqd_work_group_size

[dnkurek]

Also 896 seems relatively not a lot. It makes sense why reducing stack size would improve performance, since you are basically freeing up resources that are not used.