AMD's Radeon HD 4870 X2 - Testing the Multi-GPU Waters
by Anand Lal Shimpi & Derek Wilson on August 12, 2008 12:00 AM EST- Posted in
- GPUs
These Aren't the Sideports You're Looking For
Remember this diagram from the Radeon HD 4850/4870 review?
I do. It was one of the last block diagrams I drew for that article, and I did it at the very last minute and wasn't really happy with the final outcome. But it was necessary because of that little red box labeled CrossFire Sideport.
AMD made a huge deal out of making sure we knew about the CrossFire Sideport, promising that it meant something special for single-card, multi-GPU configurations. It also made sense that AMD would do something like this, after all the whole point of AMD's small-die strategy is to exploit the benefits of pairing multiple small GPUs. It's supposed to be more efficient than designing a single large GPU and if you're going to build your entire GPU strategy around it, you had better design your chips from the start to be used in multi-GPU environments - even more so than your competitors.
AMD wouldn't tell us much initially about the CrossFire Sideport other than it meant some very special things for CrossFire performance. We were intrigued but before we could ever get excited AMD let us know that its beloved Sideport didn't work. Here's how it would work if it were enabled:
The CrossFire Sideport is simply another high bandwidth link between the GPUs. Data can be sent between them via a PCIe switch on the board, or via the Sideport. The two aren't mutually exclusive, using the Sideport doubles the amount of GPU-to-GPU bandwidth on a single Radeon HD 4870 X2. So why disable it?
According to AMD the performance impact is negligible, while average frame rates don't see a gain every now and then you'll see a boost in minimum frame rates. There's also an issue where power consumption could go up enough that you'd run out of power on the two PCIe power connectors on the board. Board manufacturers also have to lay out the additional lanes on the graphics card connecting the two GPUs, which does increase board costs (although ever so slightly).
AMD decided that since there's relatively no performance increase yet there's an increase in power consumption and board costs that it would make more sense to leave the feature disabled.
The reference 4870 X2 design includes hardware support for the CrossFire Sideport, assuming AMD would ever want to enable it via a software update. However, there's no hardware requirement that the GPU-to-GPU connection is included on partner designs. My concern is that in an effort to reduce costs we'll see some X2s ship without the Sideport traces laid out on the PCB, and then if AMD happens to enable the feature in its drivers later on some X2 users will be left in the dark.
I pushed AMD for a firm commitment on how it was going to handle future support for Sideport and honestly, right now, it's looking like the feature will never get enabled. AMD should have never mentioned that it ever existed, especially if there was a good chance that it wouldn't be enabled. AMD (or more specifically ATI) does have a history of making a big deal of GPU features that never get used (Truform anyone?), so it's not too unexpected but still annoying.
The lack of anything special on the 4870 X2 to make the two GPUs work better together is bothersome. You would expect a company who has built its GPU philosophy on going after the high end market with multi-GPU configurations to have done something more than NVIDIA when it comes to actually shipping a multi-GPU card. AMD insists that a unified frame buffer is coming, it just needs to make economic sense first. The concern here is that NVIDIA could just as easily adopt AMD's small-die strategy going forward if AMD isn't investing more R&D dollars into enabling multi-GPU specific features than NVIDIA.
The lack of CrossFire Sideport support or any other AMD-only multi-GPU specific features reaffirms what we said in our Radeon HD 4800 launch article: AMD and NVIDIA don't really have different GPU strategies, they simply target different markets with their baseline GPU designs. NVIDIA aims at the $400 - $600 market while AMD shoots for the $200 - $300 market. And both companies have similar multi-GPU strategies, AMD simply needs to rely on its more.
Facebook
Twitter
RSS
93 Comments
View All Comments
xsilver - Tuesday, August 12, 2008 - link
Is anyone still waiting for ATI to get off their butts and fix/enhance their avivo encoder?Its incredibly fast but having the ability to encode high quality videos would be nice.
If ATI are not willing to develop it why dont they just open up the source code so that others can develop it?
TheJian - Tuesday, August 12, 2008 - link
For one, I haven't seen a 4870X2 for less than $600-650 and they don't exist now. Don't expect these to go for $550 for a while. You should print what they are being SOLD for at the time of the article.GTX260 pricing (check newegg) is off also. You can get one for $245 and it's an OC Edition (MSI)! They have quite a few at $269 retail. So where the heck do you get $300? Again quit printing suggested retail prices (or whatever they are) and PRINT ACTUAL PRICING! In the case of your GXT260 SLI price that would drop it $110 and they are both overclocked!
GTX280 isn't $450. Newegg has them for $399 if you want the cheapest, and most are $410-429. Do you guys even check the pricing before putting up your stories? You do this every time.
You state this for AMD "At the same time, AMD's literally hot GPUs have seen their prices fall; the Radeon HD 4870 is now a $270 - $280 GPU, slightly down from $299 and the Radeon HD 4850 is a $170 - $180 card. These are very slight changes in price, but at least they are in the right direction."
But you conveniently leave out that Nvidia's cards don't run $450 (GTX280 your price) or $299 (GTX260 your price). These are FAR from reality. A suggested retail doesn't matter. What matters is WHAT I WILL PAY if I buy it today! These prices haven't changed in about a week so you've had plenty of time to faq check before printing. Also the 4870 is $250/259 if you want the cheapest at newegg. So you even got the ATI pricing wrong. Newegg has some 4870X2's listed at $559+ but they won't be that by the time their auto-price-upping machine gets done with them (still higher than the $550 you state...which won't happen for a month or more likely). They'll hit $600 next week before one even sells...LOL. You need to fix the prcing in the article to reflect REALITY.
Ezareth - Wednesday, August 13, 2008 - link
I just bought a Sapphire 4870X2 from Newegg for $559 an hour ago so the pricing is correct. Obviously on launch day they are going to be sold out as people like me have been waiting for them for months now...the same was true of the 280GTXs as well. It will take a couple weeks before they become readily available and then the price will start coming down eventually to around $500.00.Aberforth - Tuesday, August 12, 2008 - link
I am studying business models of different tech companies - from the core architectures to marketing, it helps me to understand how these products actually sell, basically the crux of the tech industry.We've seen GTX 280, which was priced at $650 when it was released, but even after months of research NV comes up with a GPU that is just too large and bloated- redesign is what they actually do. Yet, it gets a good review and market hype. When you look at some of the early 260 reviews, they actually say "it's reasonably priced".
So these self proclaimed geeks write articles based on comparison- which is a childish thing to do. You cannot compare one design with another nor you can judge it's merits only looking at the performance factor. There cannot be one single outcome of a review, there are different types of customers with different requirements. So at the end of the review if someone favors either AMD or NV are biased. A unbiased review cannot contain suggestive material that hampers the customer's decision but instead it should contain information on how it affects customers with different requirements.
ZootyGray - Tuesday, August 12, 2008 - link
Thanx for that - good post.I have been thinking similar thoughts; but my ability to say it as you have done, was clouded by my own reaction of disgust at this shit-on-a-stick socalled report.
This site is seriously biased against AMD.
I used to think it was thorough testing - but issues found here, are not reported elsewhere, and are simply not experienced by users that post to forums. I expect config issues with new releases - but this site uses any excuse.
Anandtech BIAS is out of the bag.
I thought I was the only one seeing this; but I am glad to see general rejection of bullshit by many others.
Very unprofessional that we have no idea what 486-box you used to skew these results - YOUR RESULTS DON'T MATCH OTHER SITES - and they aren't shuffling dx9 and dx10 to fudge it all.
If you have to change the game midstream,
u r cheating.
.
GmTrix - Tuesday, August 12, 2008 - link
I'm surprised that the 9800 GX2 wasn't included in the benchmarks...Seeing that the 4870 X2 is ATI's most powerful single card solution and as this article states the 9800 GX2 is still Nvidias's most powerful single card solution. Not to mention they are very similarly priced and they are both dual GPU cards...
Mr Roboto - Thursday, August 14, 2008 - link
The 9800GX2 would get smoked in higher resolutions with it's much lower bandwidth and smaller frame buffer. Anything above 1600x1200 especially with AA and the GX2 would choke. That's where the 4870x2 really shines is above those resolutions. That's why.Spacecomber - Tuesday, August 12, 2008 - link
I was wondering the same thing: why leave out nvidia's dual core card? I know that it is not as powerful as the 4870 GX2, but it is nvidia's most powerful single card solution. And, the pricing seems to make it a fairly competitive option. It is selling for less than $300, which positions it well against the 4870 single core card as well as the GTX 280.It's not a big deal, but it has struck me as odd that this card hasn't been included in recent video card reviews. I know that not every available card can be rounded up for benchmarking; however, I think this is one that many consumers would be interested in knowing about, especially if they are thinking of spending around $300 on a video card.
techguy2k5 - Tuesday, August 12, 2008 - link
Mr. Anand Lal Shimpi:you have a writer working for you that has a bias towards a particular IHV. How do I know this? Because he makes it apparent in EVERY piece he writes. Derek Wilson, the constant pot-shots against ATi are pathetic. You are incapable of keeping your bias out of your articles, and therefore should not be writing. I will not read another Derek Wilson article again. In fact, I will not read another Anandtech article until something is done about this matter.
I'm not the only person aware of Mr. Wilson's bias, all of my tech enthusiast friends are aware and feel the same way. It is sad what has become of Anandtech in recent years. It used to be easy to trust Anandtech and take your writers' word on any issue, but no more. Derek Wilson is dragging this site's name through the mud.
Dismayed,
-techguy
ZootyGray - Tuesday, August 12, 2008 - link
Keys also goes overboard to find trouble.And lack of followup articles - as promised - where are they? Trouble only reported here with 780G has not been followed up as mentioned it would be - since last May. But the damage was done. Do we believe it or no?
Credibility? nah.
Bias? haha
the ntel fanboys seem to like it - or employboys.
And the ultimate recently 16 pages of praise for non-existent theoretical 2yrs out larrabeee rehash trash. Preceded by - we don't know anything about it at all. as dictated by ntel bs div.
Biased much? haha
This is dangerous market influence. How far does this go? How long has this been happening?
Violating trust is pretty sad.