Friday, May 13th 2016
More Polaris10 and Polaris11 Specifications Revealed
Industry sources revealed to TechPowerUp some pretty interesting specifications of AMD's two upcoming GPUs based on the 4th generation Graphics CoreNext "Polaris" architecture. The company is preparing a performance-segment GPU and a mainstream one. It turns out, that the performance-segment chip, which the press has been referring to as "Ellesmere," could feature 32 compute units (CUs), and not the previously thought 40.
Assuming that each CU continues to consist of 64 stream processors (SP), you're looking at an SP count of 2,048. What's more, this chip is said to offer a single-precision floating point performance of 5.5 TFLOP/s, as claimed by AMD. To put this into perspective, the company had claimed 5.2 TFLOP/s for the "Hawaii"/"Grenada" based FirePro W9100, which launched earlier this February, and that SKU featured all 2,816 SP present on the chip. So this chip is definitely faster than most "Hawaii" based SKUs.
While "Hawaii" based SKUs feature TDP of no less than 250W, the new chip has a TDP rated no higher than 150W. AMD could pull off a "single 8-pin power connector" feat like NVIDIA, with quite some headroom to spare. The chip features a 256-bit wide GDDR5/GDDR5X memory interface, and 8 GB could be its standard memory amount. The first SKUs based on this chip could feature 7 Gbps GDDR5 memory.
AMD will upgrade the feature-set to include HVEC/H.265 hardware encode/decode acceleration, DisplayPort 1.3, and HDMI 2.0a outputs.
The smaller "Polaris" chip scheduled for 2016, which the press has been referring to as "Baffin," could feature 14 compute units, working out to a stream processor count of 896. It will be a mainstream chip, succeeding the "Tobago" silicon, which drives the current R7 360 series SKUs, although it wouldn't surprise us if it outperformed bigger chips, such as the "Trinidad" based R7 370 series. This chip has its peak single-precision floating-point performance rated at 2.5 TFLOP/s. Its TDP is rated at just 50W, and it is expected to feature a 128-bit wide GDDR5 memory interface, holding 4 GB of memory.
Assuming that each CU continues to consist of 64 stream processors (SP), you're looking at an SP count of 2,048. What's more, this chip is said to offer a single-precision floating point performance of 5.5 TFLOP/s, as claimed by AMD. To put this into perspective, the company had claimed 5.2 TFLOP/s for the "Hawaii"/"Grenada" based FirePro W9100, which launched earlier this February, and that SKU featured all 2,816 SP present on the chip. So this chip is definitely faster than most "Hawaii" based SKUs.
AMD will upgrade the feature-set to include HVEC/H.265 hardware encode/decode acceleration, DisplayPort 1.3, and HDMI 2.0a outputs.
The smaller "Polaris" chip scheduled for 2016, which the press has been referring to as "Baffin," could feature 14 compute units, working out to a stream processor count of 896. It will be a mainstream chip, succeeding the "Tobago" silicon, which drives the current R7 360 series SKUs, although it wouldn't surprise us if it outperformed bigger chips, such as the "Trinidad" based R7 370 series. This chip has its peak single-precision floating-point performance rated at 2.5 TFLOP/s. Its TDP is rated at just 50W, and it is expected to feature a 128-bit wide GDDR5 memory interface, holding 4 GB of memory.
133 Comments on More Polaris10 and Polaris11 Specifications Revealed
Why produce a rushed forward (rumour) cut down Vega to fight off GTX1080? I scratch my head. Unless...
Theory 1 - AMD thought Nvidia don't yet have a DX12 answer for async so AMD didn't bother too hard. But Nvidia release (PR guff) info that 1080 is way better than expected..
Theory 2 - AMD are playing a waiting game to release cut down Vega to tease out big Pascal desktop part and then turn all the dials to 11 on a full chip Vega release to trump Nvidia at the high end.
Theory 3 - Nvidia don't have a great async answer and they rushed 1080 out, way before AMD could answer. That way Nvidia steal sales from AMD by way of having a better card out, even though it's not as good as it should be and will get smacked about by Vega.
I actually think 3 is more accurate. I think Pascal desktop will be trumped by Vega. If I'm wrong, AMD can go suck cement because if they can't build on Fiji's performance in DX12 using a smaller process node, they should be sent packing.
I do REALLY want AMD to woo me away from the 370 tho, for their own sake.
390X level performance is more than enough for 1080P gaming.
The Polaris 11 having 50W TDP and being able to match R9 280 is great for OEMs also if it ends being sold for about $150 to consumers. So, with those 2 cores and being able to have 2 SKUs available for each one, and with Polaris 11 ideal for mobiles, it might be the best strategy for AMD in order to grab a good part of the selling cake back from the green camp.
Things look really good if availability is ample from early June.
However, that 5.5 TFLOPS seems to be quite interesting, but I doubt that with the same number of CU that it'll be faster than Hawaii/Grenada.
Anyway in short, if you want to buy Nvidia cards, AMD is NOT going to make it cheaper for you. You will have to pay the full price Nvidia wants plus the extra for the founders editions.
In the end, what we have here is a duopoly in full harmony with the two competing companies staying away from each other's feet.
dream on amd
no one wants your gpu's now
Should you not be worried about that 3.5gb that was sold to you as 4gb instead?
Now im hoping for a "Shit nvidia/intel fanboys say" vid as well though, although it probably is something like "AMD sucks lelz, buy Nvidia, AMD is bankrupt and that somehow to us is a good thing lel"
449$ for 1070 (plus moar, even moar, for shortages) - cause it's "Founders Edition" (lovely, nVidia, thank you!)
699$ for 1080 (again, "Founder's Edition")
Note we are talking about prices here and now (well, short term). Who knows what will cost what when Xmas comes.Eating "1070 faster than Titan", is a real shame, although partially TPUs, for sharing that lovely piece of FUD in title.
1070 is "faster than Titan" at "certain things in VR, in certain scenarios".They needed chips suitable for PS4k is my answer.
Oh, Sony has actually PAID a couple of hundred million $ for development.+1 here.
Vega please!!!
2016 Midrange = 2015 Enthusiast
That's what has to happen for any of these releases to make sense. High overclocks on Polaris 10 and performance approaching within 10% of a 980ti at a pricepoint of $300 or less should mean we progress nicely in all ranges of price to performance.