AMD Radeon HD 7750 & Radeon HD 7770 GHz Edition Review: Evading The Price/Performance Curve
by Ryan Smith & Ganesh T S on February 15, 2012 12:01 AM EST- Posted in
- GPUs
- AMD
- HTPC
- GCN
- Radeon HD 7000
Compute Performance
Moving on from our look at gaming performance, we have our customary look at compute performance. With GCN AMD significantly overhauled their architecture in order to improve compute performance, as their long-run initiatives rely on GPU compute performance becoming far more important than it is today.
With such a move however AMD has to solve the chicken and the egg problem on their own, in this case by improving compute performance before there are really a large variety of applications ready to take advantage of it. As we’ll see AMD has certainly achieved that goal, but it raises the question of what was the tradeoff for that? We have some evidence that GCN is more efficient than VLIW5 on a per-shader basis even in games, but at the same time we can’t forget that AMD has gone from 800 SPs to 640 SPs in the move from Juniper to Cape Verde, in spite of a full node jump in fabrication technology. In the long run AMD will be better off, but I suspect we’re looking at that tradeoff today with the 7700 series.
Our first compute benchmark comes from Civilization V, which uses DirectCompute to decompress textures on the fly. Civ V includes a sub-benchmark that exclusively tests the speed of their texture decompression algorithm by repeatedly decompressing the textures required for one of the game’s leader scenes. Note that this is a DX11 DirectCompute benchmark.
Theoretically the 5770 has a 5% compute performance advantage over the 7770. In practice the 5770 doesn’t stand a chance. Even the much, much slower 7750 is ahead by 12%, meanwhile the 7770 is in a class of its own, competing with the likes of the 6870. The 7770 series still trails the GTX 560 to some degree, but once again we’re looking at the proof of just how much the GCN architecture has improved AMD’s compute performance.
Our next benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. We’re now using a development build from the version 2.0 branch, and we’ve moved on to a more complex scene that hopefully will provide a greater challenge to our GPUs.
SmallLuxGPU is another good showing for the GCN based 7700 series, with the 7770 once again moving well up the charts. This time it’s between the 6850 and 6870, and well, well ahead of the GTX 560 or any other NVIDIA video cards. Throwing in an overclock pushes things even farther, leading to the XFX BESDD tying the 6870 in this benchmark.
For our next benchmark we’re looking at AESEncryptDecrypt, an OpenCL AES encryption routine that AES encrypts/decrypts an 8K x 8K pixel square image file. The results of this benchmark are the average time to encrypt the image over a number of iterations of the AES cypher.
Under our AESEncryptDecrypt benchmark the 7770 does even better yet, this time taking the #2 spot and only losing to its overclocked self. PCIe 3.0 helps here, but as we’ve seen with the 7900 series there’s no replacement for a good compute architecture.
Finally, our last benchmark is once again looking at compute shader performance, this time through the Fluid simulation sample in the DirectX SDK. This program simulates the motion and interactions of a 16k particle fluid using a compute shader, with a choice of several different algorithms. In this case we’re using an (O)n^2 nearest neighbor method that is optimized by using shared memory to cache data.
It would appear we’ve saved the best for last, as in our fluid simulation benchmark the top three cards are all 7700 series cards. This benchmark strongly favors a well organized cache, leading to the 7700 series blowing past the 6800 series and never looking back. Even NVIDIA’s Fermi based video cards can’t keep up.
Facebook
Twitter
RSS
155 Comments
View All Comments
Hellbinder - Thursday, February 16, 2012 - link
sorry for the type the 6000 list should have started with 6900Belard - Thursday, February 16, 2012 - link
This modern new 7750 at $110 performs about the same as a 4 year old 4850... which was selling for $100 about 3 years ago. The 4800 series was more expensive to produce and drink more power.The names don't mean much anymore, they were stupid to change the stack names which were fine from the 3000~5000 series.
delirious_nomad - Thursday, February 16, 2012 - link
there are reviews out there and 7770 X-Fired smoke a single 580 and for $300 some odd dollars...I have been out of PC gaming for along time and these are going to be my cards of choice.
reasons why... I don't care what the Jones do...
I play at 1080p on an adequate LCD TV... and I don't need graphics maxed to the gills...
I have older games Half Life, Jedi Knight, Knights of the Old Republic, Diablo, Morrowind, etc etc that I still want to play and the power down features and low power usage are great boons for me.
from the X-Fire reviews I've seen so far they scale at about 2x so just double the numbers and they smoke a single 580 while using less power and running nice and quiet...
also it gives me a year or so before I build my second system and who knows what will be out then. then this gets handed down to my son and off we go. and it should be plenty fast enough for Minecraft
the only card that comes really close for me is the 560 Ti 448 Core... and one of them is more expensive and doesn't beat a 580...
here is link to techpowerups' X-Fire review... http://www.techpowerup.com/reviews/HIS/HD_7750_777...
KineticHummus - Thursday, February 16, 2012 - link
"offers performance close to NVIDIA's GeForce GTX 570."That is straight from the techpowerup link you gave, on the conclusion page. Close to gtx 570 isnt smoking the 580 which is what you stated cf 7770s will do...
CeriseCogburn - Wednesday, March 21, 2012 - link
LOL - man ... thanks.Anyway there's a triple fan GTX 580 on egg for $359.
http://www.newegg.com/Product/Product.aspx?Item=N8...
---
I said it to you so I won't get attacked but maybe the gentleman would like to reconsider in light of our helpful posts.
papapapapapapapababy - Thursday, February 16, 2012 - link
this is all nice but sorry, I got burned waaaaaay to many times by AMD BS to even care! having to wait for months for proper support, faulty drivers and mind bogglingly piss poor performance per dollar in the latest games = never again going to buy or recommend any sort of AMD graphic solution. Im going to wait for the next gen consoles to launch, and then im going to get the absolutley cheapest and most efficient nvidia solution that offers me twice the performance of whatever m$, sony chooses to put inside their lil crappy casual boxes. Just like i Always DO, but his time AMD is out of the peculation for good! se ya.BoFox - Friday, February 17, 2012 - link
The only source of this is a slideshow from AMD regarding the launch of Barts GPUs.And then AMD launched Bulldozer with a slideshow saying that it has 2B trannies. A few months after launch, AMD admitted that it was an error, saying that it has only 1.2B trannies.
I've done such an extensive performance analysis to conclude that all Barts-based GPUs (6870, 6850, and 6790) are VLIW4-based just like its Cayman cousins.
GCN appears to be around 10% more efficient than VLIW4 for games overall, but it's very hard to say exactly how much. If a 78xx card that comes out next month has very similar specs to a VLIW4-based card (heck, or a VLIW5-based one), it'd be much easier to say. Still researching on this...
CeriseCogburn - Wednesday, March 21, 2012 - link
Man it's just amazing..." And then AMD launched Bulldozer with a slideshow saying that it has 2B trannies. A few months after launch, AMD admitted that it was an error, saying that it has only 1.2B trannies. "
I see that now...
" Update: AMD originally told us Bulldozer was a 2B transistor chip. It has since told us that the 8C Bulldozer is actually 1.2B transistors. The die size is still accurate at 315mm2. "
http://www.anandtech.com/show/4955/the-bulldozer-r...
Ok, there's just no way this core transistor miscount is a mistake. The level of incompetence requried for that to be a mistake and not a PR plot is staggering.
For that reason BoFox I don't doubt what you are saying about Barts 68xx's and 6790...
I'd sure love to see that exposed far and wide as well if true. It's just amazing to me, a staggering "error" on the most basic bulldozer spec... and we're supposed to just pretend it never happened and no explanation is ever given.
Yes there's a chance you are correct Bofox on your calculations, certainly cannot put it past amd given the track record.
maree - Friday, February 17, 2012 - link
From the consumer's point of view, the 7750 is the best card which doesn't require external power connector and hence can fit in with standard case and SMPS. The 7770 makes sense for those who crossfire, esp with long idle and compares favourably against the 6950/6970, esp for somebody who plans to buy only 1 7770 now and another one later when a better deal is availableFrom AMD's point of view, the 7750 seems to be targeted at the 80% of the Market who buy Intel PC, but are envious of the graphics capabilities of the puny Llano and even tinier Brazos. The 7770 seems to be targeted at the same folks for whom the BD was targeted. If somebody was prepared to buy a product(BD/7770) which is priced closer to competition(Core-i5/Gtx560) but gets beaten in all benchmarks and is priced more than old generation(PhenomII/6850) but still loses to it in many benchmarks. In short the 7770 is a Bulldozeresque disaster.
The situation would have been much better, if they had marketed these cards as 7670 and 7750, because that is were they belong based on die/transistor size and performance. Definitely a slip-up from AMD graphics Marketing dept.
Galidou - Sunday, February 19, 2012 - link
LoL rarson, let's not get into that kind of argument with chizow, you'll end up in WW2 history of tanks pricing failure due to the fact they were not double the raw power of last gen tanks from x company vs y company... history, history, history...Be careful with Chizow's arguments, he lives in the past, nothing new to reading his comments, it's already in the books and ready for anyone that reads it to perceive it the way they want(different from an ATI or Nvidia fanboy point of view)... :P