OTOY Forums

Does double precision = double rendering speeds octane?

Ouch, crazy expensive with the same 3800 gpu cores of the previous one.
Right now there is an offer in amazon Spain for a 5700xt with 2600 Gpu cores for half price of a Radeon VII. They send to other coutries too.
https://www.amazon.es/gp/product/B07WDL ... UTF8&psc=1

I think this was mostly a hold over and a poke over at the Quadra folks.. not for us.. unless your a studio that can afford "workstation" cards.. Big Navi will be more like what we want.. looking to be like 80CU's and huge die size

First, OctaneX is software that runs on Apple Metal.
Secondly, Metal supports Apple's iOS and MacOS.
Therefore, OctaneX is not specifically designed for AMD graphics cards. Apple announced at WWDC2020 that it will be fully transferred to ARM chips two years later. At that time, it was not known if AMD graphics cards could be used, but Metal was not affected. You are discussing or paying too much attention to AMD graphics performance is meaningless!

Most of the people on here want Octane X for rendering on Macs. We can't run Nvidia cards so what else are we supposed to use? Until a viable ARM chip comes along that effectively replaces AMD GPUs, specifically for rendering that we can slot in a Mac, then we'll continue to discuss AMD GPUs, irrespective of what you think or say.

Hey guys, what do you think about the card VEGA 64 OC 8gb, it says that have 64 CU, almost similar to the big navi that will come with 80 CU.
It is interesting to see how a card of 2 years old can work almost as fast for rendering.

3 DAYS....

jenue wrote:Hey guys, what do you think about the card VEGA 64 OC 8gb, it says that have 64 CU, almost similar to the big navi that will come with 80 CU.
It is interesting to see how a card of 2 years old can work almost as fast for rendering.

3 DAYS....

Hey there! I think you shouldn't directly compare CU values to CU values, especially if we are talking bigger architectural changes to the GPU design. Its just not an accurate thing to do and I wouldn't want you to be dissapointed or overhyped

Vega is a well iterated GCN arch while Big Navi is going to be RDNA2 (The IPC alone will be different). Besides that you also have a different manufacturing process that will directly affect the clocks and power values and you quickly come to a point where its really hard to predict anything based on the available "last gen" spec sheet.

True enough.. if its written for metal. we can use any of the metal compute units whether they are AMD or maybe even a dedicated accelerator card.. think if apple gave us a rendering version of it's afterburner card! pop a few of those in.. and they could outpace any regular gpu. Not saying they will but could be interesting if they did that.

Where's that detailed blog post Goldorak promised us 'a few days' ago?

Weezer wrote:Where's that detailed blog post Goldorak promised us 'a few days' ago?

I am willing to bet it’s going to be Friday.

(I know Goldorak likes taking bets)