Which is the fastest GPU card for shader intensive stuff ?
category: general [glöplog]
NVidia (2080 TI / 2080 / 1080 TI )
or ATI (Vega) ?
For rasterisation, NVidia probably wins, but what about shader heavy stuff ? (eg : intensive raymarching used in 4k/64k intros).
or ATI (Vega) ?
For rasterisation, NVidia probably wins, but what about shader heavy stuff ? (eg : intensive raymarching used in 4k/64k intros).
I have made good experiences with gtx1080ti, but you probably won't need the ultimate heavy stuff. Gtx1050/60 will likely do the job as well ;-)
I asked because ATI VEGA cards have usually more cores than Nvidia (those cores are AFAIK are used for shader stuff)
But that probably doesn't mean they are faster...
But that probably doesn't mean they are faster...
Whatever is faster, you get double power if you buy two.
The result probably depends on your exact workload.
Looking for benchmarks, the 3dmark hall of fame has exactly one result for vega, which probably tells more about the popularity of the nvidia hardware than the actual performance differences.
Anyway, given that the result IS there says that vega64 is, in some cases at least, competitive with at least gtx 1080. Then there's rtx 2080 which, while apparently not as big performance improvement as people were hoping, is still significantly faster than the previous gen.
But these are all crazy powerful cards in any case.
Looking for benchmarks, the 3dmark hall of fame has exactly one result for vega, which probably tells more about the popularity of the nvidia hardware than the actual performance differences.
Anyway, given that the result IS there says that vega64 is, in some cases at least, competitive with at least gtx 1080. Then there's rtx 2080 which, while apparently not as big performance improvement as people were hoping, is still significantly faster than the previous gen.
But these are all crazy powerful cards in any case.
or get 2 overclocked 1070s for the same cash
In my experience GTX 1060 is a bit too slow for recent 4k and 64k intros.
If your intro is doing bandwith heavy stuff like e.g. per pixel bucketsorting for order independent transparency, AMD will yield faster results.
If your intro is doing heavy math stuff per pixel like raymarching, nvidia will yield faster results.
That's basically the main difference between both vendors, it's been like that for decades.
If your intro is doing heavy math stuff per pixel like raymarching, nvidia will yield faster results.
That's basically the main difference between both vendors, it's been like that for decades.
