NVIDIA 780a: Integrated Graphics and SLI in One
by Gary Key on May 6, 2008 12:00 AM EST- Posted in
- CPUs
Integrated Graphics
Beginning now, all new NVIDIA chipsets will ship with integrated graphics (which NVIDIA is now calling the mGPU), regardless of what market segment they are targeted at. It's a particularly bold move by NVIDIA but much appreciated given that the mGPU in all of its chipsets will receive PureVideo HD and thus can fully accelerate H.264/MPEG-2/VC1 decode streams.
While it's unlikely that many would purchase a high-end motherboard based on the NVIDIA nForce 780a SLI chipset and simply use its integrated graphics, the mGPU in the 780a is the same GPU used in the 750a, 730a, 720a and the GeForce 8200 based motherboards, so the discussion here is far reaching.
AMD 780G vs. NVIDIA 780a Graphics Architecture
AMD has built a superior Integrated graphics part this time around, both from a technical standpoint and in terms of realized performance. It isn't that AMD really went much further than NVIDIA in terms of engineering something great: they just selected a higher performance core to integrate into their chipset than NVIDIA did.
Neither AMD nor NVIDIA told us exactly how the built their interface to the system bus and system memory, but the lack of a local framebuffer does mean that fast and as low latency as possible communication with system memory are required. In both cases, the discrete GPU from which the integrated part is derived uses a 64-bit width connection to local memory. In both cases, since system memory offers a 128-bit wide but these parts make use of a wider bus to help compensate for increased latency to system memory. Increasing local (on die) cache would also help here, but since IGP solutions are as low cost as possible it doesn't seem likely that we've got loads more cache to play with.
We used 3dmark's single texture test to try to get an idea of memory bandwidth. The test largely removes computation overhead and ends up just pulling in as much data as possible as fast as possible and throwing it up on the screen. The result in MTexels/sec shows that NVIDIA has a bit of an advantage here, but the gap isn't huge. This means that performance differences will likely come down to compute power rather than bandwidth
| AMD 780G | NVIDIA nForce 780a | |
| 3DMark '06 Single Texture Fillrate | 910.6 MTexels/s | 983.4 MTexels/s |
Past here, NVIDIA and AMD integrated hardware diverge. AMD's solution is based on the RV610 graphics core. In fact, it is an RV610 core shrunk to 55nm and integrated into the Northbridge. This means we get 8 5-wide blocks of shader processors (SPs -- 40 total). In the very very worst case, we get 8 shader ops per clock (which isn't likely to happen in any real situation). Compare this to NVIDIA's G86 based 8 SP offering with a maximum of 8 shader ops per clock and we see quite a difference emerge. AMD's IGP can handle 8 vector instructions per clock and then some, while the similar code could run at 2 instructions per clock on NVIDIA hardware.
Of course, this difference isn't as decimating to NVIDIA as one might think at first blush. We must remember that NVIDIA cranks up it's shader clock to ridiculous speeds while AMD's shaders all run at core clock speed. With AMD and NVIDIA core clocks both coming in at 500MHz, NVIDIA's shader core runs at 1200MHz. In spite of the fact that AMD's part can do more operations per clock (probably averaging out to somewhere between 3x and 4x; it heavily depends on the application), NVIDIA is able to do 2.4x as many clocks per second which closes the gap a bit.
The only discrete part with 8 SPs is the GeForce 8300 which is OEM only. As of this writing, NVIDIA has not confirmed details other than core and shader speeds and the number of SPs in the part with us. They have stated that their integrated hardware is simlar to the 8400/8500 in order to optimize the benefit of Hybrid SLI, so it's possible the number of texutre and ROP units are 8 each. Of course, if half the number of SPs is "similar" to the 8400 and 8500 parts, we can't really be sure until NVIDIA confirms the details. We do know that AMD's hardware has 4 texture and 4 render outs since it is RV610. With so few SPs, and the competition sticking with 4/4 texture/render units, we suspect that this is what NVIDIA has done as well.
What is clear is that either way, AMD's hardware is more robust than NVIDIA's offering. Our performance tests reflect this, as we will soon show.
Facebook
Twitter
RSS
38 Comments
View All Comments
SiliconDoc - Monday, July 28, 2008 - link
Very glad this happens : " The good news is that ASUS has replicated several of our problems and we expect a new BIOS release shortly for use in the motherboard review. "That's what I call useful review that isn't a waste of time. Glad you have the reputation and the pull. ( one wonders what they do at Acer - I guess they wait for you guys to tell them...)
piroroadkill - Sunday, May 11, 2008 - link
I agree with the guys who are saying they need to make discrete (not discreet, jesus) GPUs consume much less power when idling, even if that means a hybridpower style segmentation of the gpu, but it should be done all on the hardware, completely transparent to the chipset and systemKGR - Saturday, May 10, 2008 - link
Maybe hybrid sli doesnt help frame rates too much , but it can make sense when nvidia integrates the Ageia pysics in gpu, then the mgpu can take the load of physics and the dgpu the graphics in hybrid mode, i dont know if it is possible but i think it is...duploxxx - Thursday, May 8, 2008 - link
Always like to read the reviews and comments from your site, but why don't you just provide proof with real data instead of a hit in the dark. You already have big parts of the data in an other review.(http://www.anandtech.com/video/showdoc.aspx?i=3232...">http://www.anandtech.com/video/showdoc.aspx?i=3232...
" Whether or not this price tag is worth the premium over the nForce 750a SLI boards is up for debate. It's not really in our opinion as we do not believe the current AMD processor series is capable of the required computational power needed to support 3-way SLI or Quad SLI configurations. This is not a knock against NVIDIA as AMD has the same problem with Quad CrossFire; it just reflects the current state of the processor offerings from AMD."
why don't you just put 9750-9850+790fx+2-3way crossfire against q6600/q9300+x38+2-3way and compare total price/performance/power but perhaps need to add a x48 board since the lack of pci-e lanes on the x38
you could do the same with lower spec P35 but then again this board has no decent feature set against current amd chipset offerings when you talk about multi gpu setup but would still be interesting to read what happens when using CF on this board against amd770
gipper - Wednesday, May 7, 2008 - link
So why isn't the 750a going to be the perfect HTPC motherboard?Aren't the two requirements for the perfect HTPC motherboard native 1080p output via HDMI and 7.1 LPCM audio on the same HDMI connector? Also, the post processing with a phenom matches the AMD 780 chip feature set.
So, I don't understand why you would say that the AMD 780 is better for HTPC's.
Or are you guys suggesting that it's best to wait for the AMD 780 refresh that includes 7.1 LPCM because the integrated graphics perform so much better?
Gary Key - Thursday, May 8, 2008 - link
I personally believe the 750a would make an excellent HTPC system if you utilize a ATX case design, might plan on using it for gaming with a discreet video card, and can afford it. The GF8300 board that just arrived is a better solution at first glance (if a uATX design and not having SLI capability is important) and compares favorably to the 780G from a price to performance viewpoint, more so than the GF8200. We will have an update on it next week.We received the 175.16 drivers right after the article went live and will have some post-processing comparisons (174.14s did not handle this right) this weekend between the two chipsets. Right now, it is a toss up in my opinion, and due to that fact, I would go NV for the multi-channel LPCM.
The Jedi - Wednesday, May 7, 2008 - link
A couple of points here:[QUOTE]
This is absolutely unacceptable and would prevent us from recommending the 780a as anything more than just another SLI motherboard. HybridPower is quite possibly the best feature for a high-end SLI user and if it won't work with 30" displays then its usefulness is severely degraded.
[/QUOTE]
I'll tell ya I use a 26" LCD TV on my desktop and it's big enough. I don't need 2560x1600. 1080p (1200p) is fine and matches the pixels on HDTV. Anything 1080p capable is completely reasonable. Just up the AA or AF if the FPS are too fast for ya. Just because Dell or Apple says Uber-users need a 30-inch LCD to be cool doesn't make it true. 24", 26", 27", these are great on today's desks. I really think a 30" LCD on my desk would be too big.
Gary Key, you da man, seriously, but proofread the article for typos.
Last point, and this goes for all of Anandtech's staff: Respect due, but seriously: NO dGPU. Call it A VIDEO CARD. Or -- A GRAPHICS CARD. Also, no calling a product from a CPU or GPU company a 'part' . Call it a chip -- or a CPU or a graphics chip, etc.
strikeback03 - Thursday, May 8, 2008 - link
I'd love to see a 2560x1600 24-26" display, the more resolution the better. If that 9 megapixel LCD weren't several thousand dollars it would be sweet.Wolfcastle - Wednesday, May 7, 2008 - link
The author should clean up the grammar a bit. Anandtech has a large audience.James5mith - Wednesday, May 7, 2008 - link
Maybe I'm just foolish here, but for the extreme overclocking crowd, I see an immediate and tangible benefit:If you happen to fry your video card while OC'ing it, you can use the onboard video as a stopgap until you get it repaired.