Our Experience with UVD and R600
It's easy to point to the presentations and say, "look, this is slightly misleading." This doesn't get across the even more convincing interactions we had with AMD personnel over the course of reviewing the HD 2900 XT. Hopefully this will help to illuminate the issue further. I'll have to put a disclaimer on this and say that I don't have a photographic memory and I forgot to bring my diary to the press briefing. This isn't normally the kind of stuff I take notes on for product launches.
At the AMD press event in Tunis, Tunisia (we still can't figure out why we went to Africa to talk about a graphics card), myself and a couple other reviewers had the good fortune of being able to see HD Decode first hand. We watched a side by side comparison of an R6xx based part and a G8x based part with interesting results that mostly pointed out the long way Intervideo and CyberLink have to go in refining their player software. We were under the impression that this was R600, though clarification was not provided.
We talked about UVD and PureVideo, and we were asked if a video decode comparison would be made against NVIDIA hardware for our HD 2900 XT launch article. Upon replying in the affirmative, we got the impression that we should expect the R600 to outperform its competition (the 8800 GTS) in video decode tests. This doesn't seem as certain without UVD, but our marketing speak filters failed to pick up on anything out of the ordinary.
Moving on, we were initially told to expect a driver to enable video decode acceleration shortly before launch. AMD also provided us with a copy of PowerDVD which was said to support AVIVO HD. We did get a newer driver near launch, but in testing it we saw very bad video decode performance. It turns out that we couldn't test video decode at all, as hardware acceleration was not enabled in our press driver. This indicates to us that, even though we saw decode acceleration working in Tunis, AMD was having more trouble than expected in the QA department.
We would expect QA issue with new UVD hardware. If we've got the same video decode hardware as a previous generation part, it seems a little odd to think that we wouldn't have a driver or software to enable the feature.
In going back through our slides for this piece, we noticed that it is explicitly stated that UVD is included on the Radeon Mobility HD 2300. This is an R5xx based part retrofitted with UVD. Of all the parts not to have this new feature, the DX9 class Mobility HD 2300 would have been the one we thought of first. Surely if this part has the new feature, the HD 2900 XT should also include it. The argument for not is, again, that higher end GPUs tend to be paired with CPUs that can handle the workload.
The reason we offer time and again for the fact that G86 and G84 include full video decode while G80 does not is that the slower parts had 6 months more development time. In NVIDIA's case, it makes complete sense that G80 would lack the more refined video decode hardware. For AMD to launch a top to bottom product line and leave a feature like this out of their highest end part is very strange.
So strange, in fact, that combined with all the other experiences we've had with this rather surreal product launch we still just can't help but doubt what we're hearing. To us, the most sensible explanation for all this is that the hardware must be there but is somehow broken - perhaps broken to the point that it will never be fixed on current silicon (i.e. via drivers) and a new spin would be required to address the problems.
In spite of what makes sense to us, AMD has indicated that UVD hardware is not at all present on R600. Their "official" statement has yet to appear, but we'll update this article when/if we get it.
It's easy to point to the presentations and say, "look, this is slightly misleading." This doesn't get across the even more convincing interactions we had with AMD personnel over the course of reviewing the HD 2900 XT. Hopefully this will help to illuminate the issue further. I'll have to put a disclaimer on this and say that I don't have a photographic memory and I forgot to bring my diary to the press briefing. This isn't normally the kind of stuff I take notes on for product launches.
At the AMD press event in Tunis, Tunisia (we still can't figure out why we went to Africa to talk about a graphics card), myself and a couple other reviewers had the good fortune of being able to see HD Decode first hand. We watched a side by side comparison of an R6xx based part and a G8x based part with interesting results that mostly pointed out the long way Intervideo and CyberLink have to go in refining their player software. We were under the impression that this was R600, though clarification was not provided.
We talked about UVD and PureVideo, and we were asked if a video decode comparison would be made against NVIDIA hardware for our HD 2900 XT launch article. Upon replying in the affirmative, we got the impression that we should expect the R600 to outperform its competition (the 8800 GTS) in video decode tests. This doesn't seem as certain without UVD, but our marketing speak filters failed to pick up on anything out of the ordinary.
Moving on, we were initially told to expect a driver to enable video decode acceleration shortly before launch. AMD also provided us with a copy of PowerDVD which was said to support AVIVO HD. We did get a newer driver near launch, but in testing it we saw very bad video decode performance. It turns out that we couldn't test video decode at all, as hardware acceleration was not enabled in our press driver. This indicates to us that, even though we saw decode acceleration working in Tunis, AMD was having more trouble than expected in the QA department.
We would expect QA issue with new UVD hardware. If we've got the same video decode hardware as a previous generation part, it seems a little odd to think that we wouldn't have a driver or software to enable the feature.
In going back through our slides for this piece, we noticed that it is explicitly stated that UVD is included on the Radeon Mobility HD 2300. This is an R5xx based part retrofitted with UVD. Of all the parts not to have this new feature, the DX9 class Mobility HD 2300 would have been the one we thought of first. Surely if this part has the new feature, the HD 2900 XT should also include it. The argument for not is, again, that higher end GPUs tend to be paired with CPUs that can handle the workload.
The reason we offer time and again for the fact that G86 and G84 include full video decode while G80 does not is that the slower parts had 6 months more development time. In NVIDIA's case, it makes complete sense that G80 would lack the more refined video decode hardware. For AMD to launch a top to bottom product line and leave a feature like this out of their highest end part is very strange.
So strange, in fact, that combined with all the other experiences we've had with this rather surreal product launch we still just can't help but doubt what we're hearing. To us, the most sensible explanation for all this is that the hardware must be there but is somehow broken - perhaps broken to the point that it will never be fixed on current silicon (i.e. via drivers) and a new spin would be required to address the problems.
In spite of what makes sense to us, AMD has indicated that UVD hardware is not at all present on R600. Their "official" statement has yet to appear, but we'll update this article when/if we get it.
Facebook
Twitter
RSS
53 Comments
View All Comments
Chunga29 - Tuesday, June 5, 2007 - link
Can't say I've really paid much attention to Driverheaven. I wonder how they got the info correct when others didn't? I also wonder if they were at the Tunisia event. The evidence that the information was not clearly made available is pretty overwhelming, though, even if it may have been present somewhere. Seems to speak to some serious issues within AMD in regards to getting info to the press and other contacts.Consider this from the Anandtech article: "Unfortunately, try as we might, we could not get UVD to work with the current drivers provided by AMD and the PowerDVD release that is supposed to enable the hardware acceleration on HD 2000 series parts."
Why on earth would AMD even provide them a version of PowerDVD to enable acceleration if H.264 accel isn't present? If they were only supposed to test MPEG and VC-1 offloading, you have to figure an email at least was sent to AMD asking why H.264 (PowerDVD in general?) wasn't working. All AMD had to do is say, "You're mistaken; here's a link. H.264 is not present on R600 but is on R610 and R630." Case closed. THG, AT, FS, TR, and plenty of others apparently all made an error and not a single one was corrected until a couple weeks later!?
7oby - Monday, June 4, 2007 - link
I do not know whether this information is correct, but my Acrobat says:created 04/25/2007
changed 05/11/2007
Chunga29 - Monday, June 4, 2007 - link
Well, it *could* be true, but dates are way too easy to fake. When did that page get published? I'd guess that the content was under revision until then.bob4432 - Monday, June 4, 2007 - link
not too long ago the a64 was kicking a$$ and the x850xtpe was king, well long in computer terms but not in normal terms. now amd/ati is at the bottom again, and this screw up seems to be following the trend between the 2/1 company.hopefully somebody will start to navigate the ship correctly because at the moment they are heading straight towards the iceburg
strikermlc - Monday, June 4, 2007 - link
This looks to simply be a not so cleverly disguised retraction on the part of Anandtech. Regardless of whether or not AMD's materials were flawlessly put together - if you make a mistake you should own up to it instead of pointing the finger at someone else.With that said, it is disappointing to see that AMD had some trouble clearly nailing down their press materials - but oversights do happen. Especially when it comes to large companies where the people putting together the presentations likely don't fully understand the material. I think the insinuation that something sinister is taking place behind the scenes is more than a little far-fetched however.
I love Anandtech and have followed you guys since the beginning - please, please don't turn into drama queens. You made an assumption and got egg on your face. It happens. Don't blow things out of proportion. Relax, and give use some more hardware reviews to geek out on. ;)
defter - Monday, June 4, 2007 - link
When it comes to large companies there are also lots of people who know the truth, in this case there were lots of people working for AMD who knew in advance that R600 will not have UVD. So why didn't anyone from AMD do anything when:
- misleading press material was released at Tunis, certainly some AMD guys knew that that material gives false impression
- AMD's partners included UVD in their lists of features
- Most of reviews during launch were claiming that R600 has UVD
Nobody from AMD came forward and cleared the matter in those cases. Only when it became apparent to the general public that R600 don't have UVD, AMD admitted it. Thus, this simply isn't a matter of a "small mistake". AMD deliberately mislead press and their partners.
Chunga29 - Monday, June 4, 2007 - link
BS! This isn't an "oversight" or "blame shifting" or any other term you care to use to try and pretend AMD didn't outright lie. It looks to me like AnandTech is coming out and saying, "Hey that article on the 2900? Well, some of the information we previously listed was incorrect, and we now know why we never could get HD video decode acceleration to work properly! So if you were interested in 2900 with AVIVO HD, stop now and don't make a purchase mistake based off faulty information!" Why the frak would AMD let AnandTech continue to try and test H.264 decode acceleration on 2900 cards if they knew damn well that it never worked in the first place!?AMD deliberately mislead a LOT of people on this subject, and it seems to me that AnandTech is right on the money here. I have seen a bunch of comments (probably by AMD funded geurilla marketing types) that basically said HD 2900 was better than 8800 GTS/GTX "because at least it gets full HD Video decode support". Looking at the slides, and seeing how AMD seemed pretty content to let the mistakes pass, I'd say AMD screwed up with the hardware on the 2900 cards. Sort of like the PureVideo stuff on GeForce 6800 never worked right, and plenty of people are still pissed about that.
The fact that several of the former ATI PR people are leaving says that something stinks in the state of AMD. Let's just hope we don't see more of this crap when Barcelona launches. "We're 20% faster clock for clock... only we're shipping at 2.0 GHz." (I'm not saying that's the case.)
I recall previous articles roasting Intel when they released shitty NetBurst processors that jacked up temperatures, underperformed, and cost more than the competition. Now AMD has processors that run at higher temperatures and perform slower, and they've got a new GPU that really jacks up the temps (60W more than 8800 GTS), costs more, and at best appears to match the NVIDIA part. Then we find out that the new AVIVO HD is really not present at all on the HD 2900 - but that's okay because... um... because... Oh! That's right! It's okay because some of you want to bury your heads in the sand and pretend that AMD isn't making mistakes right and left.
RD600? Late to market, only offers okay performance and nothing really special.
R600? VERY late to market, can't even best the competition, runs hot, costs more, and missing some touted features.
Barcelona? Late to market, and we can only hope that's the only drawback. I've got some AM2 platforms that could use a better CPU, but I won't believe AMD actually properly supports older AM2 boards until I see it. I remember a few 939 boards that didn't ever work with X2 chips, after all....
I'm all for competition, but I'm not going to go out and buy a clearly inferior product just in the hopes that the company will do better next time. Competition means I'm going to get a quad core Q6600 next month for under $300; if AMD continues to slump, I expect prices will start to rise, and then I'll be more than happy to continue running overclocked E4300/4400/6600 CPUs. I really don't need the Q6600, but given the upcoming price cuts I'll make the jump on my next upgrade. Unless of course Phenom lives up to its (silly) name.
Goty - Monday, June 4, 2007 - link
Fanboy? Yup.How is it an outright lie from AMD when numerous other sites from around the web seem to have gotten the information right?
Roy2001 - Tuesday, June 5, 2007 - link
If you are right how can AMD partners also get it wrong?Frumious1 - Monday, June 4, 2007 - link
Yeah, all the major sites somehow get it wrong and have to make corrections, and the small sites make corrections without any mention of the fact, and MANUFACTURERS print UVD on their damn boxes... but AMD certainly didn't do anything wrong!http://www.firingsquad.com/hardware/amd_ati_radeon...">FiringSquad
http://techreport.com/reviews/2007q2/radeon-hd-290...">The Tech Report updates article
http://www.tomshardware.com/2007/05/14/r600_finall...">Tom's Hardware gets it wrong
http://arstechnica.com/news.ars/post/20070514-amd-...">Ars Technica thought UVD was present
....
I could go on, but why don't you AMD fanboys that are defending them provide some links to major, REPUTABLE sites (like the above three) that got all the information correct without needing to do an update? I certainly didn't see anyone making a point about this on the launch day! But I'm just an NVIDIA/Intel fanboy, despite my X1950 XTX CrossFire setup. (Okay, granted that's running with Core 2 E6600 on 975X.)