ATI 4xxx Series Thread

Page 38 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Cookie Monster

Diamond Member
May 7, 2005
5,161
32
86
Originally posted by: BFG10K
Ha, Ha, I love pr battles. Ati says 800, but the nv says 160.
PR or not the fact remains four out of the five units are limited compared to the fifth one, unlike nVdia's SPs that can do everything. This requires a lot more massaging from the driver unlike nVidia's solution which almost dynamically balances itself.

I dont know about this one. Although nVIDIA's scalar ALUs can perform the tasks, it requires not 1 but several cycles when performing those special functions unlike ATi's where it takes 1 cycle thanks to its special ALU.

I think the 50% number comes from the missing MUL, where nVIDIA has finally learned how to take advantage of it.

And i dont think theres any shader domain in the RV770. Its not as easy to implement then some people may have led to believe. nVIDIA has been doing this since the G7x days. But heres a reason why you dont need it:

800 x 650MHz x 2 = 1.04 TFLOPs

It fits with the 1TFLOP ("terascale" engine according to AMD/ATI PR) and the 800ALUs rumours. However im just wondering how they've managed to jam all that in there. I find it weird since the ALU:TEX ratio has gone from 4:1 to 5:1. We all know that texturing is its biggest pitfall yet the rumours suggest that they've increased the ALU count by 250%!. I guess this would help in AA situations but i just wait to find out how they've managed to fit so many of them in if this is true. Maybe they aren't as costly.

2 weeks to go!
 

BFG10K

Lifer
Aug 14, 2000
22,709
2,995
126
I dont know about this one. Although nVIDIA's scalar ALUs can perform the tasks, it requires not 1 but several cycles when performing those special functions unlike ATi's where it takes 1 cycle thanks to its special ALU.
Sure, but I don't think this would be a big problem in terms of scheduling.

The point is nVidia?s SPs can almost always be doing something while up to four fifths of ATi?s could be sitting idle if they get instructions they can?t process.
 

biostud

Lifer
Feb 27, 2003
18,662
5,388
136
Originally posted by: SlowSpyder

So, let's assume all the numbers are correct, that the only difference between the cores is a mere 125MHz, but the difference in the memory is 1.6GHz (both using the same 256bit bus). With that we can pretty much conclude that either the 4850 is going to be somewhat bandwidth starved, or the 4870 will have more bandwidth then it needs. I don't see a 125MHz difference in core speed warranting a 1.6GHz change in memory speed... my gut says that the 4850 will be somewhat memroy starved, and the 4870 will pull away from the 4850 by a good margin, more then just the 125MHz core difference would account for due to having the memory bandwidth available to it that it needs vs. the 4850 that will probably have to wait on the memory at times.

The 4870 will probably scale better in higher resoultions and higher AA compared to the 4850.
 

Extelleron

Diamond Member
Dec 26, 2005
3,127
0
71
Originally posted by: manko
4850 3DMark Tests - Could be a hoax, but it doesn't seem too far off the mark.

ATI Radeon HD 4850
Q6600 @ 3GHz
3GB RAM

12494 3DMark06 1280x1024 0xAA 0xAF
11240 3DMark06 1600x1200 0xAA 0xAF
10442 3DMark06 1920x1200 0xAA 0xAF
7935 3DMark06 2560x1600 0xAA 0xAF

8487 3DMark06 1280x1024 8xAA 16xAF
6776 3DMark06 1920x1200 8xAA 16xAF
5001 3DMark06 2560x1600 8xAA 16xAF

X2789, GPU:2686, CPU:10322 - 3DMark VANTAGE Extreme PRESETS(1920x1200 4xAA 16xAF)

P6695, GPU:5989, CPU:10356 - 3DMark VANTAGE Performance PRESETS(1280x1024 0xAA 0xAF)

If that's true, the 4850 will be pretty impressive IMO.

In Vantage under Performance, I get a GPU score of 5609 with an 8800GTS 512MB @ 775/1891/2160. That would mean that HD 4850 is as powerful as a G92 with 800MHz+ core clock at its stock 625MHz.

I don't have an Extreme score to compare with, but X2789 would mean it would beat the 9800GTX there and also beat the 3870 X2 by a good margin.

 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
Originally posted by: BFG10K
I dont know about this one. Although nVIDIA's scalar ALUs can perform the tasks, it requires not 1 but several cycles when performing those special functions unlike ATi's where it takes 1 cycle thanks to its special ALU.
Sure, but I don't think this would be a big problem in terms of scheduling.

The point is nVidia?s SPs can almost always be doing something while up to four fifths of ATi?s could be sitting idle if they get instructions they can?t process.

What kind of instructions can't the 4 simple ALU's process? AFAIK the fat ALU handles things like transcendental functions, and those instructions need a separate unit on Nvidia's HW as well, they can't be processed by all ALU's. It all comes down to what instructions are most common in typical shaders, and I would think ADD and MUL are far more common than SIN/COS/EXP and the like.
 

SlowSpyder

Lifer
Jan 12, 2005
17,305
1,001
126
Originally posted by: Extelleron
HD 4870 Performance vs 9800GTX

HD 4870 is ~30-50% faster @ 1920x1200 4xAA/8xAF

http://img139.imageshack.us/im...amdslideahd4800eg3.jpg

HD 4850 Performance vs 8800GT

HD 4850 is ~25-50% faster @ 1920x1200 4xAA/8xAF

http://img137.imageshack.us/im...amdslidebhd4800mz1.jpg

No idea what so ever if those benches are real, but if so they look like very nice performing parts for the cost (assuming the cost rumors are accurate). I'm really itching to get my hands on a 4870 and see how well it can overclock... even though I have a feeling you won't need to overclock it and still have great performance.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Originally posted by: manko

ATI Radeon HD 4850
Q6600 @ 3GHz
3GB RAM

12494 3DMark06 1280x1024 0xAA 0xAF

I hope not! Obviously 3DMark06 isnt an indication of actual gaming performance but these scores are awful.

Here is how my 15 month old 8800GTS (96 SP) stacks up to 4850 (800 SP)

GT1: Return to Proxycon -- 40.81 (+16.3%) vs. 35.10
GT2: Firefly Forext -- 41.72 vs. 43.156 (+3.4%)
HDR1: Canyon Flight - 46.07 vs. 52.219 (+13.3%)
HDR2: Deep Freeze - 53.69 vs. 60.177 (+12.1%)

HD4850 is expected to be roughly 30% slower than HD4870. If HD4870 is expected to be 70% of the performance of 3870 X2, that would place HD4850 at least 40% faster than my card. Where are these numbers? 3DMark of 15000-16000 at least!
 

Foxery

Golden Member
Jan 24, 2008
1,709
0
0
Originally posted by: BFG10K
I dont know about this one. Although nVIDIA's scalar ALUs can perform the tasks, it requires not 1 but several cycles when performing those special functions unlike ATi's where it takes 1 cycle thanks to its special ALU.
Sure, but I don't think this would be a big problem in terms of scheduling.

The point is nVidia?s SPs can almost always be doing something while up to four fifths of ATi?s could be sitting idle if they get instructions they can?t process.

Are ATI's set up like a pipeline, or sitting in parrallel? I thought it was the former.

Extelleron: Delicious graphs, if true
 

HurleyBird

Platinum Member
Apr 22, 2003
2,759
1,455
136
Might be fake. Look at the smudge on the 4850 slide where the picture of the 4870 was in its slide. Also, the picture of the 4870 has 572 vertical pixels while the 4850 slide only has 563. Either it's a fake, or AMD made sloppy slides is sloppy.
 

nonameo

Diamond Member
Mar 13, 2006
5,902
2
76
I don't get why the 4870 keeps getting compared to the 9800GTX. That is THIS generation's high end. The 4870 is supposed to be the NEXT generation's high end!
 

Extelleron

Diamond Member
Dec 26, 2005
3,127
0
71
Originally posted by: nonameo
I don't get why the 4870 keeps getting compared to the 9800GTX. That is THIS generation's high end. The 4870 is supposed to be the NEXT generation's high end!

No, it's next gen midrange.

R700 is high end, not RV770. The HD 4870 will compete with the 9800GTX directly, so why not compare them?

nVidia is launching GT200 in the $450+ market, but they are not launching anything new to compete with the 4870. G92b might come sometime, but I don't think it will be enough to compete with RV770.

There is talk that nVidia may be considering lowering GTX 260 price to $399:
http://www.nordichardware.com/news,7853.html
 

Harmattan

Senior member
Oct 3, 2006
207
0
0
Originally posted by: Extelleron
Originally posted by: nonameo
I don't get why the 4870 keeps getting compared to the 9800GTX. That is THIS generation's high end. The 4870 is supposed to be the NEXT generation's high end!

No, it's next gen midrange.

R700 is high end, not RV770. The HD 4870 will compete with the 9800GTX directly, so why not compare them?

nVidia is launching GT200 in the $450+ market, but they are not launching anything new to compete with the 4870. G92b might come sometime, but I don't think it will be enough to compete with RV770.

There is talk that nVidia may be considering lowering GTX 260 price to $399:
http://www.nordichardware.com/news,7853.html

Not sure how many times this same point can be stressed until it's understood, but that is spot on. 4870 is AMDs enthusiast mid-range card, not top-of-the-line, and definitely not meant to compet with GT200. R700 (4870 X2) is AMD's top dog and the one that will be positioned against GT200.

Nvidia will most likely lower prices on G92 cards and hope they can hit the sub-$200 (below-48xx) market until it can release their next midrange... which will be the cards that compete with the single GPU 48xx's.
 

Munky

Diamond Member
Feb 5, 2005
9,372
0
76
Originally posted by: nonameo
I don't get why the 4870 keeps getting compared to the 9800GTX. That is THIS generation's high end. The 4870 is supposed to be the NEXT generation's high end!

Maybe the fact that we don't know how fast the g200 is has something to do with that. In either case, those slides look close to what I would expect from the rv770. Extra credit for scaling the graphs properly.
 

nRollo

Banned
Jan 11, 2002
10,460
0
0
Originally posted by: RussianSensation
Originally posted by: Extelleron
HD 4870 Performance vs 9800GTX

HD 4870 is ~30-50% faster @ 1920x1200 4xAA/8xAF

http://img139.imageshack.us/im...amdslideahd4800eg3.jpg

HD 4850 Performance vs 8800GT

HD 4850 is ~25-50% faster @ 1920x1200 4xAA/8xAF

http://img137.imageshack.us/im...amdslidebhd4800mz1.jpg

Much better now compared to the absolutely useless futuremark scores....

Those HD4870 numbers look pretty impressive, that will be a good card at that price point.
 

brandonb

Diamond Member
Oct 17, 2006
3,731
2
0
I currently have a Nvidia 8800 GTS 320 meg. I'm thinking of upgrading to 4850 when they are available. Can anybody speculate if that is an upgrade and how much faster a 4850 might be? Also anybody know the power requirements compared between the two cards?

I'm no longer interested in big 200 watt beasts like GT2000 or 4870's type things like that.
 

Extelleron

Diamond Member
Dec 26, 2005
3,127
0
71
Originally posted by: brandonb
I currently have a Nvidia 8800 GTS 320 meg. I'm thinking of upgrading to 4850 when they are available. Can anybody speculate if that is an upgrade and how much faster a 4850 might be? Also anybody know the power requirements compared between the two cards?

I'm no longer interested in big 200 watt beasts like GT2000 or 4870's type things like that.

The HD 4850 should probably use a bit less power than your 8800GTS.

In terms of performance, it will vary, but it will be a huge upgrade. The 8800GTS 320MB performs close to the 640MB sometimes and other times it performs much, much slower. Even in cases where it performs like the 640MB model, the 8800GT is a good 20% faster on average than the GTS and the 4850 is supposed to be ~30-50% faster than it.
 

Elfear

Diamond Member
May 30, 2004
7,126
738
126
Originally posted by: HurleyBird
Might be fake. Look at the smudge on the 4850 slide where the picture of the 4870 was in its slide. Also, the picture of the 4870 has 572 vertical pixels while the 4850 slide only has 563. Either it's a fake, or AMD made sloppy slides is sloppy.

Wombat over at XS said that the slides looked like standard NDA-type fare. He's apparently under NDA from AMD and says lots of leaked slides look like dump like maybe the creator wasn't the best with MS Office. Could still be fake but at least it lends some credibility to their authenticity.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |