AMD unofficially confirms Radeon Flagship – R9 390X - Launches at Computex

Published by

Click here to post a comment for AMD unofficially confirms Radeon Flagship – R9 390X - Launches at Computex on our message forum
data/avatar/default/avatar21.webp
A few more months of console gaming it seems. ^^
https://forums.guru3d.com/data/avatars/m/84/84507.jpg
nasty! these cards will bleed your wallet drier than at death valley.
https://forums.guru3d.com/data/avatars/m/237/237957.jpg
I wonder how efficient this new card is on power consumption? Hopefully it will not serve as an room space heater like their previous cards.
data/avatar/default/avatar14.webp
Can AMD put HBM RAM on the 290X?
https://forums.guru3d.com/data/avatars/m/242/242471.jpg
wow 8.1Tflops! I see there is also GM200 info in it with 6.9Tflops, was kinda hoping it would be ~ 8Tflops as well.. 🤓
https://forums.guru3d.com/data/avatars/m/231/231640.jpg
Shiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiit, looking great to bads its mostly speculation. Its a shame that is has to be liquid cooled, I am not a fan of liquid cooling since I have had some bad experiences 🙁
https://forums.guru3d.com/data/avatars/m/244/244064.jpg
About time!Let`s see this baby vs GM200!
data/avatar/default/avatar09.webp
Can AMD put HBM RAM on the 290X?
no way, untill they completely redesign the memory controller, and the core ( HBM have a complete different setup in the core than traditional memory )
data/avatar/default/avatar22.webp
No wonder NVidia announced the Titan X before AMD. Hopefully this card is priced at around $500 to $550, though I think it will be more than that.
https://forums.guru3d.com/data/avatars/m/202/202673.jpg
It's ULTRAAAAH-ENTHUUOOUSSIAAST...them's codewords for an undeserved $1k price tag...
https://forums.guru3d.com/data/avatars/m/123/123974.jpg
No wonder NVidia announced the Titan X before AMD. Hopefully this card is priced at around $500 to $550, though I think it will be more than that.
It will be at least double that.
https://forums.guru3d.com/data/avatars/m/137/137601.jpg
My GTX 780 Ti Classified gets 5.9 Tflops despite stock speeds (it's only little slower than GM200)! 😀
https://forums.guru3d.com/data/avatars/m/225/225084.jpg
The R9 290X has 1 Tflop more than the 980. It has 512bit bus over 980s 256bit. It has 320 GB/S vs the 980s 224 GB/S. It has 800 more Streaming processors vs the 980 as well. Them are no small differences, yet we all know the 980 is better. So don't expect those 390X numbers to be anything amazing because they aren't.
data/avatar/default/avatar21.webp
The R9 290X has 1 Tflop more than the 980. It has 512bit bus over 980s 256bit. It has 320 GB/S vs the 980s 224 GB/S. It has 800 more Streaming processors vs the 980 as well. Them are no small differences, yet we all know the 980 is better. So don't expect those 390X numbers to be anything amazing because they aren't.
Judging by those specs, its too close to make a call. 390X would need few more % (GCN optmizations) to close in at GM200, which I suppose is a realistic expectation. With that in mind, and if I really had to make a call - GM200 should be only a smidgen faster
https://forums.guru3d.com/data/avatars/m/215/215435.jpg
Comparing the amount of Streaming Processors between different vendor GPUs is like comparing apples to oranges imho
data/avatar/default/avatar15.webp
Comparing the amount of Streaming Processors between different vendor GPUs is like comparing apples to oranges imho
yes, but you can compare them and extrapolate from their own architecture its almost straight forward for GM200, not so much for 390X
https://forums.guru3d.com/data/avatars/m/242/242471.jpg
Reddoguk Well something like MonsterHunter Online benchmark can show how much Tflops matter with vram bandwidth relation, it uses direct compute for hair and that is vram bandwidth limited. If 4.8Tflops @ 980 can lag behind a Oc'ed 780gtx with ~ 5.2tflops (stock 780Ti has ~ 5.1tflops) then you know it matters when it really counts.. And 8.1 tflops stock 390x is a lot imho. GM200 might catch that with extra OC.
https://forums.guru3d.com/data/avatars/m/242/242371.jpg
Reddoguk Well something like MonsterHunter Online benchmark can show how much Tflops matter with vram bandwidth relation, it uses direct compute for hair and that is vram bandwidth limited. If 4.8Tflops @ 980 can lag behind a Oc'ed 780gtx with ~ 5.2tflops (stock 780Ti has ~ 5.1tflops) then you know it matters when it really counts.. And 8.1 tflops stock 390x is a lot imho. GM200 might catch that with extra OC.
This is probably a stupid question, but how is performance in a benchmark "where it really counts"? This GPU looks good, but then specs don't really tell you much about real world performance. I like that it's liquid cooled, as I no longer have a custom loop due to my medical condition not allowing me to maintain it properly. Not that I'm looking for a new GPU (only had this one for a few months) but this would probably be out of my price range. I look forward to learning more about this GPU.:)
data/avatar/default/avatar23.webp
Reddoguk Well something like MonsterHunter Online benchmark can show how much Tflops matter with vram bandwidth relation, it uses direct compute for hair and that is vram bandwidth limited. If 4.8Tflops @ 980 can lag behind a Oc'ed 780gtx with ~ 5.2tflops (stock 780Ti has ~ 5.1tflops) then you know it matters when it really counts.. And 8.1 tflops stock 390x is a lot imho. GM200 might catch that with extra OC.
Your method (is there a method to it?) is very confusing. Why are you using GTX 780 to make a prediction for GM200? And why you need to overclock it to make a point? Starting points are GTX 980, 290X. Extrapolation points are Titan X and 390X Starting point -> Extrapolation point GTX 980 -> Titan X 290 X -> 390X Punch in Teraflops and you'll see that Titan X over 980 is same as 390X over 290X. Meaning Titan X will keep the same advantage over 390X that 980 has over 290X. But count in some improvements on GCN and AMD should catch up 😛c1:
https://forums.guru3d.com/data/avatars/m/254/254725.jpg
I wanted to give you my money, AMD, June is so far away 🙁.