Introduction
There has been a lot of talk over the past week about the fact that the Radeon HD 2900 XT lacks UVD (Universal Video Decode) hardware for HD video decode acceleration on the GPU. We've seen sites from The Tech Report to DailyTech, and even Beyond3D reporting on the situation. While there is much discussion already out there, we would like to provide our take on things.
To recap, the situation is as follows. A huge number of press and channel partners were given the impression by AMD that their R600 GPU used to build the new Radeon HD 2900 XT contained UVD hardware. AMD's press materials are incredibly vague about the issue: they never explicitly state that the HD 2900 XT does or does not physically have UVD hardware. They do list R600 as having AVIVO HD, which they explained includes GPU offload of the entire video decode pipeline in their press materials.
In other slides and presentations, AVIVO HD and UVD have been presented as synonymous. As the entire HD 2000 line is supposed to support AVIVO HD, how could we assume that UVD hardware wasn't included on the GPU?
It seems that AMD is now trying to make it clear that R600 (the 80nm GPU for the HD 2900 XT) does not include UVD, but supports the same video decode features as the X1950. Dedicated hardware makes a big difference in entropy decoding, as we've seen with our G86 testing, but neither NVIDIA nor AMD have top end parts with full spec video decode offloading. The rationale is that high-end GPUs will be paired with CPUs capable of decoding any HD video thrown at them.
With such a huge volume of information to absorb, it seems reasonable to think that some details could get lost in the mix. But, while AMD never actually said anything untrue, they did omit details which could have easily clarified the situation. Walk with us through our experience and see for yourselves how we came to our conclusions.
There has been a lot of talk over the past week about the fact that the Radeon HD 2900 XT lacks UVD (Universal Video Decode) hardware for HD video decode acceleration on the GPU. We've seen sites from The Tech Report to DailyTech, and even Beyond3D reporting on the situation. While there is much discussion already out there, we would like to provide our take on things.
To recap, the situation is as follows. A huge number of press and channel partners were given the impression by AMD that their R600 GPU used to build the new Radeon HD 2900 XT contained UVD hardware. AMD's press materials are incredibly vague about the issue: they never explicitly state that the HD 2900 XT does or does not physically have UVD hardware. They do list R600 as having AVIVO HD, which they explained includes GPU offload of the entire video decode pipeline in their press materials.
In other slides and presentations, AVIVO HD and UVD have been presented as synonymous. As the entire HD 2000 line is supposed to support AVIVO HD, how could we assume that UVD hardware wasn't included on the GPU?
It seems that AMD is now trying to make it clear that R600 (the 80nm GPU for the HD 2900 XT) does not include UVD, but supports the same video decode features as the X1950. Dedicated hardware makes a big difference in entropy decoding, as we've seen with our G86 testing, but neither NVIDIA nor AMD have top end parts with full spec video decode offloading. The rationale is that high-end GPUs will be paired with CPUs capable of decoding any HD video thrown at them.
With such a huge volume of information to absorb, it seems reasonable to think that some details could get lost in the mix. But, while AMD never actually said anything untrue, they did omit details which could have easily clarified the situation. Walk with us through our experience and see for yourselves how we came to our conclusions.
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.)