• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

4870x2 Sideport Thread

Overslop

Overslop

Overslop

Soldato
Joined
22 Oct 2005
Posts
4,013
Location
Thailand
I've struggled to find reliable information about what exactly this is and what it would do if enabled.

Some rumours suggest it will be enabled to match/beat the newcoming nVidia GTX 295. But I don't know if there's any credence in that.

Thought I'd create a thread to try and gather as much information as possible into one thread.

Contribute if you can please! :p
 
xsp.jpg


One novel feature of the HD 4870 X2 is that it has an additional direct-GPU-to-GPU interconnect called CrossFire Sideport (XSP). The XSP offers an additional 5 GB/s interlink bandwidth between the GPUs but is not enabled at this time. Yes, you read correctly. The official reason why the XSP is disabled at this time is because that much bandwidth is not required with current applications and it will be enabled at some point in the future via driver update.
One major advantage of the XSP that I see is that transfers between the GPUs would have a lower latency. The Gen2 PCI-E bridge will certainly be fast, but it will take a short time to process the incoming data and send it out to the other GPU (<140 ns). With the XSP's point-to-point interlink this delay is eliminated. So my speculation is that AMD is working on using the XSP feature but the driver support simply doesn't work as intended at this time.
Click to expand...

http://www.techpowerup.com/reviews/Sapphire/HD_4870_X2/
 
Basically theres no real way to use it usefully. Its high speed, low latency path between the cores but current rendering technology and multi-GPU technology doesn't even work in such a way that could take advantage of it...

In theory I guess ATI could expose it to the developer for low level application specific features - I don't see it becoming very common in use tho.
 
Last edited:
Practically none - the GPUs do very little of the "talking" between themselves directly that would benefit from sideport.
 
I have to admit I'm not quite getting my head around it - but from what I can see you'd need something like hydra for sideport to be useful at all except for a very small amount of communication which doesn't suffer from going via the bridge chip... and if they did build a hydra style setup I'd have thought you'd need more like 50gig than 5gig to do it in this fashion.

But if they are planning something like that maybe they intend it to be an alternative to SFR which would give them the performance crown over the 295 in those titles easily - but most games run fine in AFR modes and would still be just as fast on SLI in 90% of cases.
 
Last edited:
It has to serve some useful purpose otherwise they wouldn't have spent extra cash implementing it into the DIE hope we can see some usefulness from it but I have to admit I'm a little skeptical also
 
I get the feeling they implemented it hoping to do something like hydra where it intelligently splits the scene up and then uses a mask to recombine - potentially giving much better scaling than crossfire/sli and then either found they didn't have enough time to get it to work before release, ran into patent issues (likely) or couldn't get it to work fast enough to be worth the difference.
 
Be nice to see it implemented but I'm not holding out for it. I even queried Steve Morris (CATALYSTCATCHER - Driver Head)/Terry Makedon (Senior Product Manager) for any sort of information about it they kept tight lipped about it and wouldn't even tell me even though I'm under NDA :(
 
Tom|Nbk said:
Be nice to see it implemented but I'm not holding out for it. I even queried Steve Morris (CATALYSTCATCHER - Driver Head)/Terry Makedon (Senior Product Manager) for any sort of information about it they kept tight lipped about it and wouldn't even tell me even though I'm under NDA :(
Click to expand...

Doesn't sound like something that would just give the X2 a minor performance boost when enabled then... and its just simply not going to give a massive boost enabled on current multi-GPU technology... so either they have something big and new or absolutely nothing at all... and personally I'm more inclind towards nothing at all.
 
Interview with AMD's Eric Demers

R3D: Now that we've started touching sensitive topics, can we find out what's up with the interconnect? Why is it disabled? Are naysayers correct in stating that it's up to the AIB whether or not the traces for it get built into the PCB, and that it's likely to not be enabled?

Eric: The sideport interconnect is fully functional in the reference design. Though we've found that with the current AFR mode of multi-GPU support, the additional bandwidth brought by the interconnect does not translate to a significant improvement in performance. However, we are continuously working on optimizing how our ATI CrossFireX™ technology scales and trying different methods, and could decide to enable the sideport if a method is found which gives better results and benefits from it.

R3D: So what happened? How did a good idea stop being a good idea? Are you disabling it due to heat/power concerns?

Eric: It's still a good idea...you have to consider that we've been working on Crossfire and on our AFR implementation for quite a few years, optimizing it. It's actually gotten to a point where it's really quite good, and difficult to improve upon performance-wise. As for heat and power concerns, that's hardly the case...the interconnect is pretty much an extra PCIE2.0 link. The power draw for that is in the realm of a few watts at most. The only reason for which it's not enabled currently is, as I've already mentioned, that it does not significantly impact performance ... and if you're not using something there's no point in enabling it.
Click to expand...
 
So basically as I said - no point with AFR or SFR, good scaling potential with a hydra type setup if they can find a way to implement it - which would take a couple of years of serious R&D.
 
They do state they're working on a new mode but I suspect by the time they do find a mode which works with it the 4870X2 will be dead and buried. I hope to eat my words..
 
Tom|Nbk said:
They do state they're working on a new mode but I suspect by the time they do find a mode which works with it the 4870X2 will be dead and buried. I hope to eat my words..
Click to expand...

I suspect your right - but even so it would be of benefit to future cards - and give nVidia something to think about :D nothing like a little competition to push the boundaries.
 
Rroff said:
I suspect your right - but even so it would be of benefit to future cards - and give nVidia something to think about :D nothing like a little competition to push the boundaries.
Click to expand...

Too right! No competition for Nvidia = not many people sitting with shiny new maxcores in their rigs, jeez knows what they`d be priced at without 4xxx :)
 
You must log in or register to reply here.
Back
Top Bottom