NVIDIA's Fermi: Architected for Tesla, 3 Billion Transistors in 2010
by Anand Lal Shimpi on September 30, 2009 12:00 AM EST- Posted in
- GPUs
Architecting Fermi: More Than 2x GT200
NVIDIA keeps referring to Fermi as a brand new architecture, while calling GT200 (and RV870) bigger versions of their predecessors with a few added features. Marginalizing the efforts required to build any multi-billion transistor chip is just silly, to an extent all of these GPUs have been significantly redesigned.
At a high level, Fermi doesn't look much different than a bigger GT200. NVIDIA is committed to its scalar architecture for the foreseeable future. In fact, its one op per clock per core philosophy comes from a basic desire to execute single threaded programs as quickly as possible. Remember, these are compute and graphics chips. NVIDIA sees no benefit in building a 16-wide or 5-wide core as the basis of its architectures, although we may see a bit more flexibility at the core level in the future.
Despite the similarities, large parts of the architecture have evolved. The redesign happened at low as the core level. NVIDIA used to call these SPs (Streaming Processors), now they call them CUDA Cores, I’m going to call them cores.
All of the processing done at the core level is now to IEEE spec. That’s IEEE-754 2008 for floating point math (same as RV870/5870) and full 32-bit for integers. In the past 32-bit integer multiplies had to be emulated, the hardware could only do 24-bit integer muls. That silliness is now gone. Fused Multiply Add is also included. The goal was to avoid doing any cheesy tricks to implement math. Everything should be industry standards compliant and give you the results that you’d expect.
Double precision floating point (FP64) performance is improved tremendously. Peak 64-bit FP execution rate is now 1/2 of 32-bit FP, it used to be 1/8 (AMD's is 1/5). Wow.
NVIDIA isn’t disclosing clock speeds yet, so we don’t know exactly what that rate is yet.
In G80 and GT200 NVIDIA grouped eight cores into what it called an SM. With Fermi, you get 32 cores per SM.
The high end single-GPU Fermi configuration will have 16 SMs. That’s fewer SMs than GT200, but more cores. 512 to be exact. Fermi has more than twice the core count of the GeForce GTX 285.
Fermi | GT200 | G80 | |
Cores | 512 | 240 | 128 |
Memory Interface | 384-bit GDDR5 | 512-bit GDDR3 | 384-bit GDDR3 |
In addition to the cores, each SM has a Special Function Unit (SFU) used for transcendental math and interpolation. In GT200 this SFU had two pipelines, in Fermi it has four. While NVIDIA increased general math horsepower by 4x per SM, SFU resources only doubled.
The infamous missing MUL has been pulled out of the SFU, we shouldn’t have to quote peak single and dual-issue arithmetic rates any longer for NVIDIA GPUs.
NVIDIA organizes these SMs into TPCs, but the exact hierarchy isn’t being disclosed today. With the launch's Tesla focus we also don't know specific on ROPs, texture filtering or anything else related to 3D graphics. Boo.
A Real Cache Hierarchy
Each SM in GT200 had 16KB of shared memory that could be used by all of the cores. This wasn’t a cache, but rather software managed memory. The application would have to knowingly move data in and out of it. The benefit here is predictability, you always know if something is in shared memory because you put it there. The downside is it doesn’t work so well if the application isn’t very predictable.
Branch heavy applications and many of the general purpose compute applications that NVIDIA is going after need a real cache. So with Fermi at 40nm, NVIDIA gave them a real cache.
Attached to each SM is 64KB of configurable memory. It can be partitioned as 16KB/48KB or 48KB/16KB; one partition is shared memory, the other partition is an L1 cache. The 16KB minimum partition means that applications written for GT200 that require 16KB of shared memory will still work just fine on Fermi. If your app prefers shared memory, it gets 3x the space in Fermi. If your application could really benefit from a cache, Fermi now delivers that as well. GT200 did have an L1 texture cache (one per TPC), but the cache was mostly useless when the GPU ran in compute mode.
The entire chip shares a 768KB L2 cache. The result is a reduced penalty for doing an atomic memory op, Fermi is 5 - 20x faster here than GT200.
415 Comments
View All Comments
yacoub - Thursday, October 1, 2009 - link
uh-oh, boys, he's foaming at the mouth. time to put him down.SiliconDoc - Thursday, October 1, 2009 - link
Ah, another coward defeated. No surprise.yacoub - Wednesday, September 30, 2009 - link
"The motivation behind AMD's "sweet spot" strategy wasn't just die size, it was price."LOL, no it wasn't. Not when everyone, even Anandtech staff, anticipated the pricing for the two Cypress chips to be closer to $199 and $259, not the $299 and $399 they MSRP'd at.
This return to high GPU prices is disheartening, particularly in this economy. We had better prices for cutting edge GPUs two years ago at the peak of the economic bubble. Today in the midst of the burst, they're coming out with high-priced chips again. But that's okay, they'll have to come down when they don't get enough sales.
SiliconDoc - Thursday, October 1, 2009 - link
It was fun for half a year as the red fans were strung along with the pricing fantasy here.Now of course, well the bitter disappointment, not as fast as expected and much more costly. "low yields" - you know, that problem that makles ati "smaller dies" price like "big green monsters" (that have good yields on the GT300).
--
But, no "nothing is wrong, this is great!" Anyone not agreeing is "a problem". A paid agent, too, of that evil money bloated you know who.
the zorro - Thursday, October 1, 2009 - link
silicon duck, please take a valium i'm worried about you.SiliconDoc - Thursday, October 1, 2009 - link
Another lie, no worry, you're no physician, but I am SiliconDoc, so grab your gallon red water bottle reserve for your overheating ati card and bend over and self administer you enema, as usual.araczynski - Wednesday, September 30, 2009 - link
sounds like ati will win the bang for the buck war this time as well. at least it makes the choice easier for me.marc1000 - Wednesday, September 30, 2009 - link
Some time ago I heard that the nex gen of consoles would run DX11 (Playstation2 and Xbox were DX7, PS3 and X360 DX9. So PS4 and X720 could perfectly be DX11). If this is the case, we are about to see new consoles with really awesome graphics - and then the GPU race would need to start over to more and more performance.Do you guys have any news on those new consoles development? It could complete the figure in the new GPU articles this year.
Penti - Friday, October 2, 2009 - link
I think you mean DX9 class hardware, PS3 who has zero DX9 support and XBOX360 has DX9c class support but a console specific version. PS3 was using OpenGL ES 1.0 with shaders and other feature from 2.0 as it was release pre OpenGL ES 2.0 spec. The game engines don't need the DX API. It doesn't matter to game content developers any way.Xbox was actually DirectX 8.1 equivalent. As said next gen consoles are years away. Larrabee and fermi will have been long out by then.
haukionkannel - Thursday, October 1, 2009 - link
Rumours says that next generation consoles are releaset around 2013-2014...But who can say...