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
If Pascal is the performance winner because it is blazingly fast - that is what matters. I would love to see AMD push the clocks higher because in the past they could do it well. Perhaps without HBM to hold it back Polaris will come in at a mid point but with overclocking push it above it's fighting weight.
The problem was everybody was stuck on 28nm for so many years. Since you could squeeze more transistors onto a die, Nvidia simply decided to scale back on compute power (they already have other cards for this) and repurpose transistors to do actual GPU work. This is how they were able to pretty much run circles around AMD wrt efficiency.
Bulldozer was supposed to be their counter strategy to Intel's lead in the silicon fabrication technology. So instead of competing against Intel using the same approach, that is building big cores and prioritizing single thread performance, they decided to do something different, aka bulldozer, which evidently didn't succeed. That is behind us now. AMD is releasing Zen this year. A Zen core will have double the integer resources and quadruple the floating point resources of a Bulldozer derived CPU core, in addition to high bandwidth low-latency L3 cache. Expect Zen to compete with intel's kaby Lake.
With regard to Polaris, I'm quite glad that AMD chose to release their mainstream cards first. I'm not planning on spending more than $350 CAD (~$250 USD) for a graphics card, and I think AMD's 480/480X are going to be great performers in that price category.
Based on AMD's recent track record, I think it would be incredibly optimistic to expect their new designs to be on par with Intel and Nvidia.
Anyways, BS is usual for any company, NVIDIA and AMD are notorious for doing them.
Oh yes, one thing to say, I really hated, I mean REALLY hated that they decided with the new name "Fury" (Same goes for Titan). Very confusing, I wish they'll do better on naming scheme department, maybe calling Polaris 480, 470 and Vega 490 and maybe 495. Or heck, Polaris can be 470 and 460, while 480 and 490 for Vega.
I'm pretty sure only trolls and AMD fanboys try to make "Pascal 10X Maxwell" applicable across the board, rather than a the sole example it is intended for.