The GF100 Recap

NVIDIA first unveiled its GF100 (then called Fermi) architecture last September. If you've read our Fermi and GF100 architecture articles, you can skip this part. Otherwise, here's a quick refresher on how this clock ticks.

First, let’s refresh the basics. NVIDIA’s GeForce GTX 480 and 470 are based on the GF100 chip, the gaming version of what was originally introduced last September as Fermi. GF100 goes into GeForces and Fermi goes into Tesla cards. But fundamentally the two chips are the same.

At a high level, GF100 just looks like a bigger GT200, however a lot has changed. It starts at the front end. Prior to GF100 NVIDIA had a large unified front end that handled all thread scheduling for the chip, setup, rasterization and z-culling. Here’s the diagram we made for GT200 showing that:

NVIDIA's GT200

The grey boxes up top were shared by all of the compute clusters in the chip below. In GF100, the majority of that unified front end is chopped up and moved further down the pipeline. With the exception of the thread scheduling engine, everything else decreases in size, increases in quantity and moves down closer to the execution hardware. It makes sense. The larger these chips get, the harder it is to have big unified blocks feeding everything.

In the old days NVIDIA took a bunch of cores, gave them a cache, some shared memory and a couple of special function units and called the whole construct a Streaming Multiprocessor (SM). The GT200 took three of these SMs, added texture units and an L1 texture cache (as well as some scheduling hardware) and called it a Texture/Processor Cluster. The old GeForce GTX 280 had 10 of these TPCs and that’s what made up the execution engine of the GPU.

NVIDIA's GF100

Click to Enlarge

With GF100, the TPC is gone. It’s now a Graphics Processing Cluster (GPC) and is made up of much larger SMs. Each SM now has 32 cores and there are four SMs per GPC. Each GPC gets its own raster engine, instead of the entire chip sharing a larger front end. There are four GPCs on a GF100 (however no GF100 shipping today has all SMs enabled in order to improve yield).

Each SM also has what NVIDIA is calling a PolyMorph engine. This engine is responsible for all geometry execution and hardware tessellation, something NVIDIA expects to be well used in DX11 and future games. NV30 (GeForce FX 5800) and GT200 (GeForce GTX 280), the geometry performance of NVIDIA’s hardware only increases roughly 3x in performance. Meanwhile the shader performance of their cards increased by over 150x. Compared just to GT200, GF100 has 8x the geometry performance of GT200, and NVIDIA tells us this is something they have measured in their labs. This is where NVIDIA hopes to have the advantage over AMD, assuming game developers do scale up geometry and tessellation use as much as NVIDIA is counting on.

NVIDIA also clocks the chip much differently than before. In the GT200 days we had a core clock, a shader clock and a memory clock. The core clock is almost completely out of the picture now. Only the ROPs and L2 cache operate on a separate clock domain. Everything else runs at a derivative of the shader clock. The execution hardware runs at the full shader clock speed, while the texture units, PolyMorph and Raster engines all run at 1/2 shader clock speed.

Cores and Memory

While we’re looking at GF100 today through gaming colored glasses, NVIDIA is also trying to build an army of GPU compute cards. In serving that master, the GF100’s architecture also differs tremendously from its predecessors.

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).


GT200 SM

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 GF100 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.


GF100 SM

NVIDIA’s GT200 had a 16KB shared memory in each SM. This didn’t function as a cache, it was software managed memory. GF100 increases the size to 64KB but it can operate as a real L1 cache now. In order to maintain compatibility with CUDA applications written for G80/GT200 the 64KB can be configured as 16/48 or 48/16 shared memory/L1 cache. GT200 did have a 12KB L1 texture cache but that was mostly useless for CUDA applications. That cache still remains intact for graphics operations. All four GPCs share a large 768KB L2 cache.

Each SM has four texture units, each capable of 1 texture address and 4 texture sample ops. We have more texture sampling units but fewer texture addressing units in GF100 vs. GT200. All texture hardware runs at 1/2 shader clock and not core clock.

 NVIDIA Architecture Comparison G80 G92 GT200 GF100 GF100 Full*
Streaming Processors per TPC/GPC 16 16 24 128 128
Texture Address Units per TPC/GPC 4 8 8 16 16
Texture Filtering Units per TPC/GPC 8 8 8 64 64
Total SPs 128 128 240 480 512
Total Texture Address Units 32 64 80 60 64
Total Texture Filtering Units 64 64 80 240 256
*There are currently no full implementations of GF100, the column to the left is the GTX 480

 

Last but not least, this brings us to the ROPs. The ROPs have been reorganized, there are now 48 of them in 6 parttions of 8, and a 64bit memory channel serving each partition. The ROPs now share the L2 cache with the rest of GF100, while under GT200 they had their own L2 cache. Each ROP can do 1 regular 32bit pixel per clock, 1 FP16 pixel over 2 clocks, or 1 FP32 pixel over 4 clocks, giving the GF100 the ability to retire 48 regular pixels per clock. The ROPs are clocked together with the L2 cache.

Threads and Scheduling

While NVIDIA’s G80 didn’t start out as a compute chip, GF100/Fermi were clearly built with general purpose compute in mind from the start. Previous architectures required that all SMs in the chip worked on the same kernel (function/program/loop) at the same time. If the kernel wasn’t wide enough to occupy all execution hardware, that hardware went idle, and efficiency dropped as a result. Remember these chips are only powerful when they’re operating near 100% utilization.

In this generation the scheduler can execute threads from multiple kernels in parallel, which allowed NVIDIA to scale the number of cores in the chip without decreasing efficiency.


GT200 (left) vs. GF100 (right)

With a more compute leaning focus, GF100 also improves switch time between GPU and CUDA mode by a factor of 10x. It’s now fast enough to switch back and forth between modes multiple times within a single frame, which should allow for more elaborate GPU accelerated physics.

NVIDIA’s GT200 was a thread monster. The chip supported over 30,000 threads in flight. With GF100, NVIDIA scaled that number down to roughly 24K as it found that the chips weren’t thread bound but rather memory bound. In order to accommodate the larger shared memory per SM, max thread count went down.

  GF100 GT200 G80
Max Threads in Flight 24576 30720 12288

 

NVIDIA groups 32 threads into a unit called a warp (taken from the looming term warp, referring to a group of parallel threads). In GT200 and G80, half of a warp was issued to an SM every clock cycle. In other words, it takes two clocks to issue a full 32 threads to a single SM.

In previous architectures, the SM dispatch logic was closely coupled to the execution hardware. If you sent threads to the SFU, the entire SM couldn't issue new instructions until those instructions were done executing. If the only execution units in use were in your SFUs, the vast majority of your SM in GT200/G80 went unused. That's terrible for efficiency.

Fermi fixes this. There are two independent dispatch units at the front end of each SM in Fermi. These units are completely decoupled from the rest of the SM. Each dispatch unit can select and issue half of a warp every clock cycle. The threads can be from different warps in order to optimize the chance of finding independent operations.

There's a full crossbar between the dispatch units and the execution hardware in the SM. Each unit can dispatch threads to any group of units within the SM (with some limitations).

The inflexibility of NVIDIA's threading architecture is that every thread in the warp must be executing the same instruction at the same time. If they are, then you get full utilization of your resources. If they aren't, then some units go idle.

A single SM can execute:

GF100 FP32 FP64 INT SFU LD/ST
Ops per clock 32 16 32 4 16

 

If you're executing FP64 instructions the entire SM can only run at 16 ops per clock. You can't dual issue FP64 and SFU operations.

The good news is that the SFU doesn't tie up the entire SM anymore. One dispatch unit can send 16 threads to the array of cores, while another can send 16 threads to the SFU. After two clocks, the dispatchers are free to send another pair of half-warps out again. As I mentioned before, in GT200/G80 the entire SM was tied up for a full 8 cycles after an SFU issue.

The flexibility is nice, or rather, the inflexibility of GT200/G80 was horrible for efficiency and Fermi fixes that.

Meet the GTX 480 and GTX 470, Cont Odds & Ends: ECC & NVIDIA Surround Missing
POST A COMMENT

197 Comments

View All Comments

  • Sunburn74 - Sunday, March 28, 2010 - link

    It is absolutely ridiculous. Like having a buzzsaw in your case.
    http://www.hardocp.com/article/2010/03/26/nvidia_f...">http://www.hardocp.com/article/2010/03/26/nvidia_f...
    Reply
  • Belard - Sunday, March 28, 2010 - link

    AMD already stated they are not reducing their pricing any time soon. This is because their line-up is far healthier than Nvidia.

    They know (and we should know) that the $500/$350 price for the new GeForce 4 cards are not going to stick. There is only some many thousands of cards available for the next 3~5 months. The supply will run dry in about 1-3 weeks I bet. We're going to see the pricing shoot up close to $600 for the GF480, the fanboyz will be willing to pay that price.

    The 5850 was supposed to be a $250 card, we see how well that worked out. While the price did settle around $300, the 5850 was still a better value than the $370~400 GeForce 285 card as it was far faster and run cooler, etc. The 5870 is typically faster than the $500 GeForce 295 - for $100 less. ATI has no reason to lower their pricing.

    The GeForce 265~295 cards are already being phased out, too slow, cost too much.

    So nVidia has nothing for the sub $300 market... nothing. Only the GF-250 has any value but a tad expensive as it should be $100 since its still a DX10 re-badged 9800GTX.

    So when ATI feels any pressure from Nvidia, they can easily drop their prices. It costs $5000 per wafer, no matter how many chip dies are on it. It may be costing nVidia $150~200 per chip while for AMD, they could be paying $20~35 per chip used in the 5800/5900s.
    Then you add the costs for memory, the PCB, parts, cooling system etc.

    It is very easy for AMD to drop $50 per GPU and they'd still make a profit while Nvidia sells their geForce 400 cards at a loss or no profit.

    When ATI sells the 5830 for $190~200, 5850 at $250 and 5870 at $325~350 would help sales and keep nVidia at bay.

    We'll see...
    Reply
  • SirKronan - Sunday, March 28, 2010 - link

    I would've liked to see 5850's in crossfire thrown into this mix. I know you don't have time to test them all, but I think that's the key competitor against the 480 when it comes to bang/buck. I would think the 5850's in crossfire could handily beat the 295 and the 480, all while consuming less power. I believe there may have been another site that did it, but with this excellent and very thorough review done here, it would've been even that little tiniest bit sweeter to have a 5850 crossfire line on their graphs.

    Regardless, thanks for the informative review!
    Reply
  • B3an - Saturday, March 27, 2010 - link

    This review and any other pages on this site are not working in Firefox... they have been reported as attack pages.

    "This web page at www.anandtech.com has been reported as an attack page and has been blocked based on your security preferences."

    This is why FF with all default settings, and just using adlock.
    Could it be trouble with the ads again using a malware/virus?
    Reply
  • Ryan Smith - Sunday, March 28, 2010 - link

    We know. It's being worked on. Reply
  • Sunburn74 - Saturday, March 27, 2010 - link

    The 5850 idles inthe mid 30's but it also does absolutely nothing to stay cool, operating at about 20% max fan speed. Under load it may go up to 30% fan speed, but rarely ever breaks the 40% mark.

    What are the approximate idle and load fan speeds for both the gtx 480 and 470? I guess I'm asking this to understand just how much extra cooling room is innately available. Are these card working at max capacity to keep cool or is there thermal/fan headroom there to be had?
    Reply
  • Ryan Smith - Saturday, March 27, 2010 - link

    I don't have that data on-hand (we don't record fan speeds), but it's something that I should be able to easily grab at a later time. Reply
  • Lemonjellow - Saturday, March 27, 2010 - link

    For some reason Chrome is flagging this article as a malicious sight... Oddness... Possibly got a bad advertisement... Reply
  • NJoy - Saturday, March 27, 2010 - link

    well, Charlie was semi-accurate, but quite right =))What a hot chick... I mean, literately hot. Way too hot Reply
  • WiNandLeGeNd - Saturday, March 27, 2010 - link

    Looking back at the data, I realized that power consumption is for system total. Guru3d measured the power consumption of the card itself and reported a max of 263W, so roughly 21 A. I think my 850W will do just fine since each PCI-X con has 20A each. Reply

Log in

Don't have an account? Sign up now