x1600xt Performance

stelleg151

Senior member
Sep 2, 2004
822
0
0
Is anyone else as dissapointed in the x1600xt results as I am?

According to the Anandtech benches, it is outperformed most of the time by a lower clocked, 8 pipeline card: the famous 6600gt. The 6600gt is how old?? and costs how much??? I feel that although ATI came out with a competitive product for Nvidia's NEW 7 series, they have really been blown out of the water by Nvidia's 6600 series. First came the x700 series, that didnt last, then came the x800gt and all of those odd variants, but none seem to really stand out against the 6600gt. Now the even newer generation, coming out more than a year after the 6600gt (maybe closer to 2??) still can barely hold up against the 6600gt.

As someone who doesnt always have the money for the highest end, I am pretty dissapointed with the mid-range from this gen of ATI. What happened to the new gen mid-range beating out previous flagship cards, like the 6600gt did to the 9800pro???

Anyways, basically just wanted to know if others feel the same way.
 

compgeek89

Golden Member
Dec 11, 2004
1,860
0
76
It is QUITE dissappointing. Not only is it only 6600 GT competition, its priced as 6800 competition.
 

doofer

Junior Member
Sep 30, 2005
10
0
0
Yep... tend to agree and for those of us still running an AGP rig, the stories even worse! Like how many months before any AGP varients appear? (if ever!). Hence why I decided last night to go for an X800pro. Found one for a more than resonable price, less than half what they were going for 6 months ago. They even overclock quite nicely from what I've read, hence why the card fan purchased as well.

I also can't afford the bleeding edge, so have to wait for prices to settle.

Think ATI has a rocky road ahead. I know serveral long time ATI fans who've moved over to the Nvidia camp. Life... who ever said it was easy!
 

TStep

Platinum Member
Feb 16, 2003
2,460
10
81
Yes, very large gap in performance between 1600-1800 line. Leading me to believe that ATI may be leaving some room in their lineup for 1800 cores that didn't cut the mustard for functional quads, clockspeed, etc.
 

johnnqq

Golden Member
May 30, 2005
1,659
0
0
i was expecting the x600xt to be similar to the 6600gt- it was way better than the last gen's top card. well this is bull...i hope the 7600gt will give some kind of a performance jump. this generation really does suck compared to the last one.
 

stelleg151

Senior member
Sep 2, 2004
822
0
0
Originally posted by: TStep
Yes, very large gap in performance between 1600-1800 line. Leading me to believe that ATI may be leaving some room in their lineup for 1800 cores that didn't cut the mustard for functional quads, clockspeed, etc.

Or possibly making a x1700 lineup with crap x1800 cores. I hope they do that, to reduce the number of cards with the x1800 label (we dont want another 20 card x8** fiasco).
 

Pr0d1gy

Diamond Member
Jan 30, 2005
7,774
0
76
Originally posted by: compgeek89
Originally posted by: Pr0d1gy
X800GTO blows away the 6600gt...

for a price, its $30 more min

Wow, $30 more to get something capable of more than double the speed. I think I would spend $75 more in that situation without hesitation.
 

DRavisher

Senior member
Aug 3, 2005
202
0
0
Isn't this in line with the previous generation too? The x600 and x700 werent much competition for the 6600GT either, and the X1600 will probably not compete well with a future 7600 series either. ATi seems to favour adding lots of lower perf. parts to the highend productline (x800, and now x1800 series). I am betting we will see some x1800 parts to fill the gap.
 

lifeguard1999

Platinum Member
Jul 3, 2000
2,323
1
0
We'll see of the x1600 is as bad as it looks right now once it has become available. Who knows, ATI may change it up some.
 

crazydingo

Golden Member
May 15, 2005
1,134
0
0
X1600 suffers from the memory bandwidth due to its 128bit memory, otherwise the gpu is fine especially in SM3:

Today we saw that the Shader Model 3.0 support was really done for good: just look how the RADEON X1600 XT manages to leave behind a far more expensive and enhanced GeForce 7800 GTX during dynamic branching and pixel shaders 3.0 processing.
 

mooncancook

Platinum Member
May 28, 2003
2,874
50
91
if X1600 sucks so bad, I'm hoping ATI will be pulling the 9500Pro trick once again.
 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
The x1600 does suck in terms of competitive performance, but it's an interesting card becase it hints of the future.

From b3d:
"Regular participants of our forums may well have been aware of some reasonably obscure numbering schemes for many months that were used to describe parts of the performance characteristics, such as 16-1-1-1 for R520, 4-1-3-2 for RV530 and 4-1-1-1 for RV515, but until now the exact meaning of these weren't known. With the specifications for each of the chips we can now derive that the first number equates to the number of ROP's, the second the number of texture units per ROP, the third the number of "shader pipes" per ROP, and finally the Z/Stencil multiplier per ROP - with these figures in mind, we'll let you consider the ramifications for parts in the R5xx series that are still to come..."

The same leaked slides that showed the numbers for the r520, rv530, and rv515 also described the r580 as a 16-1-3-1 design. Which means it will have 16 ROP's with 3 shaders per ROP, giving it a total of 48 pixel shaders. Now there's something to think about :D
 

stelleg151

Senior member
Sep 2, 2004
822
0
0
Originally posted by: munky
The x1600 does suck in terms of competitive performance, but it's an interesting card becase it hints of the future.

From b3d:
"Regular participants of our forums may well have been aware of some reasonably obscure numbering schemes for many months that were used to describe parts of the performance characteristics, such as 16-1-1-1 for R520, 4-1-3-2 for RV530 and 4-1-1-1 for RV515, but until now the exact meaning of these weren't known. With the specifications for each of the chips we can now derive that the first number equates to the number of ROP's, the second the number of texture units per ROP, the third the number of "shader pipes" per ROP, and finally the Z/Stencil multiplier per ROP - with these figures in mind, we'll let you consider the ramifications for parts in the R5xx series that are still to come..."

The same leaked slides that showed the numbers for the r520, rv530, and rv515 also described the r580 as a 16-1-3-1 design. Which means it will have 16 ROP's with 3 shaders per ROP, giving it a total of 48 pixel shaders. Now there's something to think about :D

That is interesting, but if its any indicator for what those 3 shaders per ROP will do, then I for one am not too excited for R580(considering those rumors of 16-1-3-1 are true.)
 

stelleg151

Senior member
Sep 2, 2004
822
0
0
Originally posted by: crazydingo
X1600 suffers from the memory bandwidth due to its 128bit memory, otherwise the gpu is fine especially in SM3:

We learned in the 6600gt vs 9800pro that thats not all that matters. The thing that bothers me here is the x1600 vs 6600gt. 6600gt also has 128bit memory and 8 pipes and it still beats the x1600xt. The 6600gt has become what it has become (3rd? most used card on Valve survey) because of its bang for the buck. Thats what I was looking for from this gen of mid range: bang for the buck, perhaps replacing the x800 and 6800 series with a new, faster, cheaper to manufacture, more efficient midrange card. Instead we got a more expensive, slower card. Thus my discontent.

The x1800 series was brought out to compete with the 7800 series, and since NVidia still hadnt come out with a new gen midrange card, this was ATI's chance for this gen to beat NVidia to the chase and make a card that would be the 6600gt of this generation. But all the x1600 does is barely compete with NVidias LAST generation midrange. Thus my discontent.
 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
Originally posted by: stelleg151
Originally posted by: munky
The x1600 does suck in terms of competitive performance, but it's an interesting card becase it hints of the future.

From b3d:
"Regular participants of our forums may well have been aware of some reasonably obscure numbering schemes for many months that were used to describe parts of the performance characteristics, such as 16-1-1-1 for R520, 4-1-3-2 for RV530 and 4-1-1-1 for RV515, but until now the exact meaning of these weren't known. With the specifications for each of the chips we can now derive that the first number equates to the number of ROP's, the second the number of texture units per ROP, the third the number of "shader pipes" per ROP, and finally the Z/Stencil multiplier per ROP - with these figures in mind, we'll let you consider the ramifications for parts in the R5xx series that are still to come..."

The same leaked slides that showed the numbers for the r520, rv530, and rv515 also described the r580 as a 16-1-3-1 design. Which means it will have 16 ROP's with 3 shaders per ROP, giving it a total of 48 pixel shaders. Now there's something to think about :D

That is interesting, but if its any indicator for what those 3 shaders per ROP will do, then I for one am not too excited for R580(considering those rumors of 16-1-3-1 are true.)

The 3 shaders per ROP is a good thing. The problem with the x1600 is that it only has 4 ROP's, and consequently only 4 texture units. In a card with 16 each, things would be completely different.
 

stelleg151

Senior member
Sep 2, 2004
822
0
0
Originally posted by: munky
The 3 shaders per ROP is a good thing. The problem with the x1600 is that it only has 4 ROP's, and consequently only 4 texture units. In a card with 16 each, things would be completely different.

Oh wow, wait im an idiot. Sorry I didnt read that report, or even register those numbers. So really the the x1600xt is a 4 pipe part with 3 rop?? Thats great to hear, so I now I see what you were saying before about future potential.

Forgive my intellectual sluggishness.
 

hans030390

Diamond Member
Feb 3, 2005
7,326
2
76
for the price it sucks, but its not a bad card.

what's really disappointing is that it was supposed to score around 6800-6800gt marks in 3Dmark05, which it didn't...so that $250 price mark isn't looking to nice...
 

Pete

Diamond Member
Oct 10, 1999
4,953
0
0
No, no, stelleg, the X1600 is a four ROP GPU with three pixel shaders per ROP. So, 4 ROPs and 4 * 3 = 12 PS (as well as 5 VS).
 

xtknight

Elite Member
Oct 15, 2004
12,974
0
71
Originally posted by: Pete
No, no, stelleg, the X1600 is a four ROP GPU with three pixel shaders per ROP. So, 4 ROPs and 4 * 3 = 12 PS (as well as 5 VS).

So essentially the ROPs are bottlenecking...what? The pixel shader pipeline? Isn't that a very idiotic thing to do? What a waste of power? Isn't that like putting some artificial cap on your hard drive speed when it could theoretically go 50MB/sec, or am I completely misunderstanding it?
 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
Originally posted by: xtknight
Originally posted by: Pete
No, no, stelleg, the X1600 is a four ROP GPU with three pixel shaders per ROP. So, 4 ROPs and 4 * 3 = 12 PS (as well as 5 VS).

So essentially the ROPs are bottlenecking...what? The pixel shader pipeline? Isn't that a very idiotic thing to do? What a waste of power? Isn't that like putting some artificial cap on your hard drive speed when it could theoretically go 50MB/sec, or am I completely misunderstanding it?

I dont think it's the actual ROP's limiting, but the 4 texture units. The 6600gt only has 4 ROP's and it doesnt seem to be bottlenecked. But when it comes to texturing, the x1600 basically acts like a 4 pipe card. And that does seem like a stupid thing to do, becasue I feel it would have been much more competitive with 12 texture units.
 

stelleg151

Senior member
Sep 2, 2004
822
0
0
I think I get it now, thanks munky. Although I'm unsure how you could get ahold of stats like that? How did you find out that the 6600gt has 4 ROPs? (also what does ROP stand for?)

Thanks a lot, I thought I knew my stuff when it came to vid cards but im being learned reallll good.
 

DRavisher

Senior member
Aug 3, 2005
202
0
0
Originally posted by: munky
The same leaked slides that showed the numbers for the r520, rv530, and rv515 also described the r580 as a 16-1-3-1 design. Which means it will have 16 ROP's with 3 shaders per ROP, giving it a total of 48 pixel shaders. Now there's something to think about :D

Wouldn't 48 pixel shaders be like 500million transistors or something? Or maybe they have sverely crippled each pipeline or something? I just don't see ATi delivering such a part.
 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
It is hard to believe that the r580 wil have 48 pixel shaders, but those leaked slides held up for the current Ati cards, so it's reasonable to expect them to hold up for the r580. This was discussed at B3D for months before the new Ati cards were launched.