4870x2 XSP Sideport feature poised to counter rumored 280gx2

B

BigCactus

Gawd
Joined
Jul 2, 2008
Messages
743
We've already seen benches where the 4870x2 is comparable and already oustsa GTX280 sli set up in a few instances. So with the slim possibility of a 280gx2, it would seem as though that the 4870x2 would need a tad more performance under it's belt to out perform the 280gx2. Well that performance is actually already on the card...it's just not being utilized by the current drivers. This "novel" feature is the XSP. Consider it a built in "Step up Program." ;)

"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."

Source: http://www.techpowerup.com/reviews/Sapphire/HD_4870_X2/
 
so this link would be the microstudder elimination that was originally talked about with the 4870x2?
 
I was thinking it's possible for todays games, it creates overhead that actually hurts performance. Future games might need the boost it could deliver and the gains would offset any overhead enabling it might cause.
 
tissimo said:
so this link would be the microstudder elimination that was originally talked about with the 4870x2?
Click to expand...

Yes, the sideport enables the GPUs to communicate with each other, effectively eliminating any unsynced frames.
 
tissimo said:
so this link would be the microstudder elimination that was originally talked about with the 4870x2?
Click to expand...


can we just shut up about micro stuttering? i bet you 20 bucks 75 peorcent of everyone here has no idea what theyre even referring to.
 
I have also heard that they are not going to require the custom boards to have it though, that would seem to me that it didn't work as they thought. its only on the reference cards right now
 
"Microstuttering" is the new "Bottleneck" in terms of widely used terms that people use without understanding wtf they are talking about.

Anyway, more bandwidth between the GPUs isn't going to make a big difference. By the time it does make a difference, the 4870X2 will be old news anyway.
 
I don't wanna hear about the labor pains... just give me the baby .

BigCactus said:
We've already seen benches where the 4870x2 is comparable and already oustsa GTX280 sli set up in a few instances. So with the slim possibility of a 280gx2, it would seem as though that the 4870x2 would need a tad more performance under it's belt to out perform the 280gx2. Well that performance is actually already on the card...it's just not being utilized by the current drivers. This "novel" feature is the XSP. Consider it a built in "Step up Program." ;)

"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."

Source: http://www.techpowerup.com/reviews/Sapphire/HD_4870_X2/
Click to expand...
 
AmongTheChosenX said:
can we just shut up about micro stuttering? i bet you 20 bucks 75 peorcent of everyone here has no idea what theyre even referring to.
Click to expand...

You can't ignore measurable facts with the fallacy called "Argumentum ad numerum"
(Appeal to majority)

Even when the majority op people belived the earth was flat, it was infact round...

You fail...
 
i get your point, but everyone is making a mountain out of a molehill with micro stuttering... for goodness sake.
 
theres very little to begin with. i ahvent even ehard of micro stuttering till a few months ago... and i've been into computer since the radeon 9600 series and a little while before
 
AmongTheChosenX said:
theres very little to begin with. i ahvent even ehard of micro stuttering till a few months ago... and i've been into computer since the radeon 9600 series and a little while before
Click to expand...

That you havn't heard of it before means squat...another fallacy...the world dosn't revolve around your knowlegde.
The issue is real, it can be meassured...hell you can even see it in videos :rolleyes:

And saying "and i've been into computer since the radeon 9600 series and a little while before"...makes me laugh. :D

I was into computers before this, but how long you (or I) have been into computers is yet another irrelevant fallacy.

3 x fail = hat-trick...congrats.
 
Well, if Crysis can't use the bandwidth right now, what makes them think something is magically going to start needing it, they are getting their asses handed to them in Crysis right now. I'm begining to think it's bogus that it will improve performance at all, or it's just plain broken.

If it is true that they disabled it just to sand bag nVidia, how do you like paying a premium for a intentionally crippled card. That's frames you could have been enjoying for how ever long it takes for them to decide they are ready to give you what you paid for.
 
http://www.anandtech.com/video/showdoc.aspx?i=3372&p=3

So why isn't this feature working for the folks whom have picked these up. Read up. I know for one thing, if I were to pickup a 4870X2, wait however months for a 280GX2 variant to appear and then have a magically improved card suddenly I would not be impressed that I should have had that all along. So just what is the truth regarding sideport exactly?


 
Atech said:
...your eyes are worthless as "evidence".
Click to expand...

Arguments about microstuttering specifically aside, that makes very little sense when talking about graphics, an entirely visual medium.
 
Hmmm.

So anyone wanna take bets on how long it takes for someone to find the location in the drivers that's keeping this locked up and breaking it?
 
Atech said:
Care to run FRAPS and confirm this, as your eyes are worthless as "evidence".
Here is some REAL evidence:
http://www.pcgameshardware.com/aid,...es_micro_stuttering_on_the_Radeon_HD_4870_X2/

What is up with people and fallacies today? :eek:
Click to expand...

a quick play through of Crysis

Frame, Time (ms)
1, 0.000
2, 49.646
3, 102.039
4, 155.181
5, 202.662
6, 253.703
7, 324.830
8, 375.051
9, 422.995
10, 470.629
11, 521.823
12, 573.016
13, 622.636
14, 676.204
15, 765.499
16, 815.874
17, 864.960
18, 914.017
19, 964.047
20, 1014.496
21, 1067.056
22, 1118.742
23, 1167.424
24, 1219.598
25, 1272.663
26, 1324.092
27, 1372.593
28, 1421.731
29, 1474.627
30, 1525.802
31, 1572.716
32, 1620.740
33, 1666.353
34, 1714.091
35, 1796.860
36, 1844.414
37, 1896.561
38, 1950.143
39, 2003.122
40, 2051.616
41, 2098.554
42, 2151.660
43, 2203.281
44, 2253.847
45, 2299.740
46, 2345.863
47, 2393.667
48, 2440.401
49, 2487.175
50, 2533.083
 
Guys can we talk about sideport, I think there are 10 or more threads on microstuttering!


 
Direct quote from the Anandtech link given by Wind, previously:

"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."

From that tidbit, I highly doubt the HD4870 X2's Sideport would be enabled to counter possible faster options from Nvidia in the future. And even if it was, you're talking about negligible increases in performance (if that)...
 
So it's actually not even a feature available on current production cards, that we know of, due to the traces not bieng there at all. But I suspect there is something off about it, since Crysis could most assuredly use the extra performance if it existed. Anyone see a 4970x2 in the future?
 
MartinX said:
Arguments about microstuttering specifically aside, that makes very little sense when talking about graphics, an entirely visual medium.
Click to expand...

Can your eyes meassure frame times in ms or even FSP...YES/NO?

merlin704 said:
I bet you believe in global warming also...:rolleyes:
Click to expand...

Ah....Ad Hominem fallacy...sorry that you failed...feel free to make an ass of yourself again :rolleyes:
(I could link to genmay, to post where I distance myself from the notion of global warmning as a manmade phenomenon, but why waste anymore of my time on you useless post?)

eric66 said:
lol unfortunately he uses his eyes to play games :rolleyes:
Click to expand...

Can your eyes meassure frame times in ms or even FSP...YES/NO? :rolleyes:

I love post based on ignorance...not.:rolleyes:
 
Atech said:
Can your eyes meassure frame times in ms or even FSP...YES/NO?
Click to expand...

He has a point. Sure, you can't feasibly confirm or deny the existence of microstuttering with just your eyes. If we were arguing about the existence of microstuttering your rebuttal would be appropriate. However the point that was being made is that if a user's eyes are incapable of perceiving microstuttering when it happens, then microstuttering is not detrimental for that user, and thus not an issue anyway. Nice strawman, btw.
 
Kaldskryke said:
He has a point. Sure, you can't feasibly confirm or deny the existence of microstuttering with just your eyes. If we were arguing about the existence of microstuttering your rebuttal would be appropriate. However the point that was being made is that if a user's eyes are incapable of perceiving microstuttering when it happens, then microstuttering is not detrimental for that user, and thus not an issue anyway. Nice strawman, btw.
Click to expand...

It's not a fallacy, why do you think we have sites like Hard (doing benches).
But microstutter is in the same ballpark as the "shimmering" issue.
Many people didn't notice it...untill someone made them aware af the issue..and then they couldn't ignore it.
Same thing with cue marks
Untill made aware of them, you don't "percive" them...but they are there.
 
Atech said:
Can your eyes meassure frame times in ms or even FSP...YES/NO?



Ah....Ad Hominem fallacy...sorry that you failed...feel free to make an ass of yourself again :rolleyes:
(I could link to genmay, to post where I distance myself from the notion of global warmning as a manmade phenomenon, but why waste anymore of my time on you useless post?)



Can your eyes meassure frame times in ms or even FSP...YES/NO? :rolleyes:

I love post based on ignorance...not.:rolleyes:
Click to expand...

The only person here making an ass of themself, is you.

As for microstutter, I don't notice it in my 3 rigs. 2x8800GTX Ultra, 2x GTX280 and 2x 4870x2.
 
merlin704 said:
The only person here making an ass of themself, is you.

As for microstutter, I don't notice it in my 3 rigs. 2x8800GTX Ultra, 2x GTX280 and 2x 4870x2.
Click to expand...

That you don't NOTICE it, dosn't mean it isn't there...fail again :rolleyes:
 
Atech said:
That you don't NOTICE it, dosn't mean it isn't there...fail again :rolleyes:
Click to expand...

But if he doesn't notice it, it doesn't matter, does it?

You come flaming everyone here with your 'evidence' in the form of some graphs, but what's the relevance of those if you're measuring something that's not visually perceivable?

I have a 9800GX2, I should have micro-stutters too, but so far I haven't managed to notice anything, I even tried a few tests and paid extra attention, but none whatsoever.

If it's technically measurable like in your stupid little graphs, I sure believe you. It's there. But I cannot see it, so I don't care. This has nothing to do with ignorance whatsoever. Why should I change my hardware for something I cannot see? Why do these micro-stutter crusaders keep trying to tell everyone not to buy hardware setups that include micro-stutter, when it's actually of no relevance?
 
Guys he says it exists and it matters, just accept it. Who cares if you can't see it, it's there and you need to be extremely worried or the evil Microstutter fairy will get you in your sleep! :rolleyes:
 
Xeanor said:
But if he doesn't notice it, it doesn't matter, does it?

You come flaming everyone here with your 'evidence' in the form of some graphs, but what's the relevance of those if you're measuring something that's not visually perceivable?

I have a 9800GX2, I should have micro-stutters too, but so far I haven't managed to notice anything, I even tried a few tests and paid extra attention, but none whatsoever.

If it's technically measurable like in your stupid little graphs, I sure believe you. It's there. But I cannot see it, so I don't care. This has nothing to do with ignorance whatsoever. Why should I change my hardware for something I cannot see? Why do these micro-stutter crusaders keep trying to tell everyone not to buy hardware setups that include micro-stutter, when it's actually of no relevance?
Click to expand...

Try to run the level, like the video here:
http://www.pcgameshardware.com/aid,...es_micro_stuttering_on_the_Radeon_HD_4870_X2/

And report back.
 
myndcrime said:
a quick play through of Crysis

Frame, Time (ms)
1, 0.000
2, 49.646
3, 102.039
4, 155.181
5, 202.662
6, 253.703
7, 324.830
8, 375.051
9, 422.995
10, 470.629
11, 521.823
12, 573.016
13, 622.636
14, 676.204
15, 765.499
16, 815.874
17, 864.960
18, 914.017
19, 964.047
20, 1014.496
21, 1067.056
22, 1118.742
23, 1167.424
24, 1219.598
25, 1272.663
26, 1324.092
27, 1372.593
28, 1421.731
29, 1474.627
30, 1525.802
31, 1572.716
32, 1620.740
33, 1666.353
34, 1714.091
35, 1796.860
36, 1844.414
37, 1896.561
38, 1950.143
39, 2003.122
40, 2051.616
41, 2098.554
42, 2151.660
43, 2203.281
44, 2253.847
45, 2299.740
46, 2345.863
47, 2393.667
48, 2440.401
49, 2487.175
50, 2533.083
Click to expand...

The same level as in the video...or?
 
Atech said:
That you don't NOTICE it, dosn't mean it isn't there...fail again :rolleyes:
Click to expand...

if a tree falls in the woods and there is no one there to hear it does it make any noise?

A visual defect that can not be perceived is irrelevant. If they have reduced micro stuttering to below the threshold of perception then it is by any reasonable person "fixed"
 
Atech said:
...fail again :rolleyes:
Click to expand...

Wow, if I just write "Fail" at the end of every statement, does that mean that I automatically win the argument???

Reminds me of snot-nosed kids way back in the eight grade who would punctuate their arguments with such amazingly cogent words as "Moded" and "Boo-Yaaa."
 
You must log in or register to reply here.
Back
Top