Go Back   nV News Forums > Hardware Forums > Benchmarking And Overclocking

Newegg Daily Deals

Reply
 
Thread Tools
Old 10-02-03, 06:39 PM   #121
Moose
Cheating is for losers!!!
 
Join Date: Jul 2002
Posts: 241
Default

Quote:
Originally posted by StealthHawk
I'm not attacking the driver team. Just pointing out the coincidences in benchmark scores between ATI and NVIDIA. You don't think it's odd that pre-optimization NVIDIA gets its ass handed to it and post-optimization scores are virtually in a deadlock with ATI? It should be apparent that NVIDIA sets ATI's scores as a target and "optimizes" as much as needed to get to that goal.

I have noticed this for quite a while now. The steady decline of Nvidia's image quality as the benches have gone up is obvious as well.

No smoke and mirrors at least.. those are turned off along with lighting and fog.

Moose is offline   Reply With Quote
Old 10-02-03, 06:49 PM   #122
The Baron
Guest
 
Posts: n/a
Default

Quote:
Originally posted by demonic
R360 uses Low-K?

Link pls
It doesn't, RV360 does.`
  Reply With Quote
Old 10-02-03, 06:50 PM   #123
demonic
Registered User
 
Join Date: Feb 2003
Posts: 10
Default

Quote:
Originally posted by The Baron
It doesn't, RV360 does.`
Yep Yep..

cos the way Hellbinder is talking.. hes gonna get his core and mem to 500!!
demonic is offline   Reply With Quote
Old 10-02-03, 07:05 PM   #124
AnteP
Nordic Nerd
 
AnteP's Avatar
 
Join Date: Dec 2002
Location: Sweden, Lund
Posts: 552
Send a message via ICQ to AnteP
Default

Quote:
Originally posted by Hellbinder
ROFL!!!!

You are trying to suggest that the ATi card looks like a 6 year old put it together.. ROFLMAO!!!! Thats just plain *Beep*. Maybe you should look at a few more pics of the actual ATi model.

The R360 and PCB are designed around overclocking. It even uses Low-K. the overclocking potential of the R360 both core and Ram are *Much* higher than the iceQ.
the RV360 is using low k, NOT the R360
notice the V

if they were actually using low k I'd doubt they'd need such a big ass heatsink

for the record: the XT memory gets hot AS HELL
the heatsink on the back of the board will burn the skin off your fingers if you're not carefull I know from experience
not quite as hot as the 5800 Ultra but almost

I think they should have made the heatsink on the back a bit thicker and with fins
I bet some 3rd party manufacturer will do that, I hope so at least

and oh yeah, one more thing IceQ roxxors your boxxors =) very efficient and very silent
a XT with that cooling and a batter heatsink at the back and I'm ready to go

the good part is that IceQ can be bought as a standalone cooling set and fits just fine on the XT

Last edited by AnteP; 10-02-03 at 07:08 PM.
AnteP is offline   Reply With Quote
Old 10-02-03, 07:46 PM   #125
Ghola
Registered User
 
Ghola's Avatar
 
Join Date: Feb 2003
Location: Winnipeg
Posts: 132
Default

Quote:
Originally posted by jAkUp
its still lower iq, and probably even lower iq since these cheating drivers, not to mention it still gets owned by 30fps in techdemo 5

when is punkbuster gonna detect driver cheats??

Punkbuster can't detect sh!t, CD maybe that'll do it!
lastest cheats, ogc/mygot/new dets. hehe
Ghola is offline   Reply With Quote
Old 10-02-03, 10:19 PM   #126
mikechai
Aegophile
 
mikechai's Avatar
 
Join Date: Jul 2002
Location: Malaysia
Posts: 700
Default

Does Hellbinder really think R360 using low-k or just typing too fast?
__________________
P4 1.6A @ 2382 MHz | 149x16@1.625V | Asus P4S533 |
256MBx2 Samsung PC2700 | MSI GF2gts 32MB DDR
mikechai is offline   Reply With Quote
Old 10-02-03, 10:25 PM   #127
theultimo
DoD + Sharepoint=Headache
 
theultimo's Avatar
 
Join Date: Sep 2003
Location: Iraq
Posts: 667
Default

Well, I personally bellieve the driver team went downhill from when the team left to ATi. From 10-20-30-40..sorta there was modest improvements in the optimization of the cards. Considering that the NV3x is hard to optimize for, and also considering all they've done to the 40 series is mostly driver cheats, 50's might be the same way, but the 51.75 was interesting that they actually worked on features instead of cheating for once...
__________________
MAIN MACHINE: Macbook Pro Currently being used in Clamshell Mode | Core Duo 2.16 | 2GB DDR-2 667 | Mobility Radeon X1600 256MB (Underclocked version) | Windows Vista/OS X | 19' Viewsonic 1912wb
UMPC: Samsung Q1 | 512MB DDR-2 | 40GB | Xp Tablet PC 2005 | Intel GMA 900
XBOX Live Gamertag: theultimos
theultimo is offline   Reply With Quote
Old 10-02-03, 10:58 PM   #128
jimmyjames123
Registered User
 
Join Date: May 2003
Posts: 665
Default

Quote:
It sure sounds to me like the reordered instructions in an optimal way for NV3x.
The "mixed mode" setting on HL2 from those preliminary benchmarks is a bit puzzling. Look at the results carefully.

http://www.techreport.com/etc/2003q3/valve/index.x?pg=2

The 5600 Ultra appears to LOSE 2-3fps (sadly, about 20%) when moving from DirectX 9.0 mode to the mixed mode (something that Beyond3d and other websites apparently didn't seem to catch?). And the 5200 Ultra only gains about 1-2 fps at most when moving from DirectX 9.0 mode to the mixed mode. The 5900 Ultra benefits much more from the move to mixed mode. Clearly there are some issues, probably relating to both hardware and software. However, it certainly doesn't seem that HL2's mixed mode is particularly well optimized for the FX cards, at least at the moment with the currently available drivers!

On a side note, I'm sure that some people noticed that these preliminary benchmarks were done at 1024x768x32, with no AA and no AF. Anand got some negative flack from people for showing data using these settings (including that "new" set of HL2 data from an unnamed source), some of whom were perfectly content with the prelim HL2 numbers at similar settings. Go figure.

Last edited by jimmyjames123; 10-02-03 at 11:05 PM.
jimmyjames123 is offline   Reply With Quote

Old 10-02-03, 11:12 PM   #129
Bopple
Registered User
 
Join Date: Mar 2003
Posts: 208
Default

Think it's due to different scales of the two graphs and inaccurate drawing of FX5600 graph bars.
According to numeric results done by some sites, 5600 in mixed-mode was 1~2 frame faster than in full-precision mode.
__________________
Handsome fighter never loses battle.
Bopple is offline   Reply With Quote
Old 10-02-03, 11:15 PM   #130
SlyBoots
Registered User
 
SlyBoots's Avatar
 
Join Date: Jul 2002
Location: La Grande, OR
Posts: 339
Question

IIRC the mixed mode is for 5900 only, the 5600/5200 will only run DX8
SlyBoots is offline   Reply With Quote
Old 10-02-03, 11:17 PM   #131
jimmyjames123
Registered User
 
Join Date: May 2003
Posts: 665
Default

Do you have a source that has confirmed that? Pretty big error to make in a presentation with graphs, don't you agree?

Even with a 1-2 fps increase for the 5600 Ultra, really doesn't seem like a very well optimized coding path. Definitely an issue with both hardware and software.
jimmyjames123 is offline   Reply With Quote
Old 10-02-03, 11:19 PM   #132
jimmyjames123
Registered User
 
Join Date: May 2003
Posts: 665
Default

Quote:
IIRC the mixed mode is for 5900 only, the 5600/5200 will only run DX8
I would think mixed mode was done for all the FX cards in general (that's why they showed them all on the graph). All the FX cards (including the 5600 and 5200) can technically run the game in mixed mode or DirectX 9.0, just not very well at all (at least with the currently available drivers).
jimmyjames123 is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 09:01 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright 1998 - 2014, nV News.