SPCC: Testrun of Lc0 Bad Gyal 9P XL finished

[pohl4711]

NN-testrun of Lc0 0.26.3 Bad Gyal 9P XL finished

https://www.sp-cc.de

(Perhaps you have to clear your browsercache or reload the website)

[dkappe]

NN-testrun of Lc0 0.26.3 Bad Gyal 9P XL finished

https://www.sp-cc.de

(Perhaps you have to clear your browsercache or reload the website)

I suspect there’s something wrong with your afterburner slowdown. BGXL 9p is much stronger than the original 9.

[pohl4711]

NN-testrun of Lc0 0.26.3 Bad Gyal 9P XL finished

https://www.sp-cc.de

(Perhaps you have to clear your browsercache or reload the website)

I suspect there’s something wrong with your afterburner slowdown. BGXL 9p is much stronger than the original 9.

Afterburner just reduces the MHz of the GPU. Compensates the using of Cuda 11 by Lc0, which makes Lc0 faster. There is not "something wrong". Afterburner displays the MHz of GPU and I measured the speed of Lc0 with and without Cuda 11 and with and without using the Afterburner-slowdown.

[dkappe]

NN-testrun of Lc0 0.26.3 Bad Gyal 9P XL finished

https://www.sp-cc.de

(Perhaps you have to clear your browsercache or reload the website)

I suspect there’s something wrong with your afterburner slowdown. BGXL 9p is much stronger than the original 9.

Afterburner just reduces the MHz of the GPU. Compensates the using of Cuda 11 by Lc0, which makes Lc0 faster. There is not "something wrong". Afterburner displays the MHz of GPU and I measured the speed of Lc0 with and without Cuda 11 and with and without using the Afterburner-slowdown.

Well, several different testers have found 9p to be 30-40 elo stronger than 9, and clearly stronger in h2h. Otherwise I can’t explain your results.