@cmuratori naïve port of DX11 -> DX12 renderer: whoops, 2x slower. Try to optimize a bunch: about same perf.
-
-
Replying to @aras_p
@cmuratori with decently optimized drivers (DX11 - years of tuning), it's hard to beat them if your hi-level code is "DX9/11/GL style"2 replies 2 retweets 6 likes -
Replying to @aras_p
@aras_p@cmuratori@matiasgoldberg Full ack. For D3D12/Vulkan, you really need to optimize API usage in your app to beat the D3D11 driver.2 replies 0 retweets 4 likes -
Replying to @NIV_Anteru
@NIV_Anteru@aras_p@matiasgoldberg And nobody thinks this is a problem??? How many devs are going to know how to do that?4 replies 0 retweets 4 likes -
Replying to @cmuratori
@cmuratori@NIV_Anteru@matiasgoldberg Yeah, all 17 of them will know what to do :) It might be a problem for teaching.3 replies 1 retweet 2 likes -
Replying to @aras_p
@aras_p@cmuratori@matiasgoldberg I don't expect Vulkan/D3D12 to be taught at universities anytime soon.3 replies 0 retweets 0 likes -
Replying to @NIV_Anteru
@NIV_Anteru@aras_p@cmuratori@matiasgoldberg: woah, hold your horses there. Let's get memory cache related issues be taught at univ first.1 reply 0 retweets 2 likes -
Replying to @Zavie
@Zavie@NIV_Anteru@aras_p@matiasgoldberg Or we just ride it out until Intel finishes making memory accesses free :)2 replies 0 retweets 0 likes -
Replying to @cmuratori
@cmuratori@NIV_Anteru@aras_p@matiasgoldberg: I wouldn't hold my breath. Also, it's about time parallel programming is taught too.1 reply 0 retweets 1 like -
Replying to @Zavie
@Zavie@cmuratori@NIV_Anteru@matiasgoldberg That's easy. Never take any locks == lockless. Right?2 replies 0 retweets 0 likes
@aras_p @Zavie @NIV_Anteru @matiasgoldberg "This program is guaranteed to work for at least one ordering of the instructions."
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.