azonenberg changed the topic of #scopehal to: ngscopeclient, libscopehal, and libscopeprotocols development and testing | https://github.com/ngscopeclient/scopehal-apps | Logs: https://libera.irclog.whitequark.org/scopehal
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 268 seconds]
Degi_ is now known as Degi
Degi has quit [Quit: ZNC 1.8.2+deb2build5 - https://znc.in]
Degi has joined #scopehal
<_whitenotifier-9> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal/compare/1422ffd75f07...7156faad7a49
<_whitenotifier-9> [scopehal] azonenberg 7156faa - ThunderScope: Added support for 50 ohm coupling
<_whitenotifier-9> [scopehal-apps] azonenberg pushed 3 commits to master [+0/-0/±5] https://github.com/ngscopeclient/scopehal-apps/compare/812625e3bb29...d72278ab9a11
<_whitenotifier-9> [scopehal-apps] azonenberg 7fa276f - Changed working buffer in render shader to int, only use float for final output
<_whitenotifier-9> [scopehal-apps] azonenberg 227c26b - Switched rendering loop to use atomics instead of lots of barriers for ~30% speedup on RTX 2080 Ti
<_whitenotifier-9> [scopehal-apps] azonenberg d72278a - Moved a bunch of variables out of shared memory to reduce memory BW, changed threads per block to 128 rows. Now 77% speedup vs original on a 2080 Ti
<azonenberg> So i just did a bunch of optimization to the rendering shader and got a 77% speedup on nvidia and, i think, even more significant on intel iGPU
<_whitenotifier-9> [scopehal] azonenberg pushed 1 commit to master [+1/-0/±3] https://github.com/ngscopeclient/scopehal/compare/7156faad7a49...968163dfa224
<_whitenotifier-9> [scopehal] azonenberg 968163d - ThunderScope: removed very slow clip detection, started work on GPU clip detection
<_whitenotifier-9> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://github.com/ngscopeclient/scopehal-apps/compare/d72278ab9a11...2fcecb1c8efd
<_whitenotifier-9> [scopehal-apps] azonenberg 2fcecb1 - Updated to latest scopehal
<d1b2> <azonenberg> With the new optimizations, 40 WFM/s and 60FPS for 2M points on an Intel CML GT2 iGPU, 320 Msps / 2.56 Gbps
<d1b2> <azonenberg> (with a thunderscope)
<d1b2> <azonenberg> And on a Quadro RTX 3000 dGPU on the same laptop, 10M points * 25 WFM/s * 4 channels = full 1 Gsps streaming