Sign in with
Sign up | Sign in
Your question

Kyro II misrepresented!!!

Tags:
  • Graphics Cards
  • Graphics
Last response: in Graphics & Displays
Share
Anonymous
a b U Graphics card
December 21, 2001 12:03:08 AM

Ok: recently i've seen the Kyro II appear multiple times in Toms Hardware articles as well as at other publications. This is a good thing. However, i am appalled at the results it gets.

As a faithful Kyro II user i feel that i should point out that THE KYRO II SCORES MUCH BETTER THAN THG HAS REPORTED.

I would like to point out that in a recent article THG reported the Kyro II to have achieved 84.9 fps on Quake III (1024x768 etc..). First off i would like to say that i am perplexed by the lack of further details about the settings. Are we to assume that this everything is on High Quality?

Now: The score i have been getting at this resolution (1024x768x32) is 101 fps this is unoverclocked. I admit: i have a 1466mhz AXP system. But that does not account for a 15% fps drop? (I would try this at 1200mhz but i can't since my multiplier is not unlocked; i support fsb overclocking).

So i ran some tests @ 1024x768etc... MAX settings

GL Driver: Default
GL Extensions: On
Video Mode: 1024x768
Color Depth: 32 bit
Fullscreen: On
Lighting: Lightmap
Geometric Detail: High
Texture Detail: Max
Texture Quality: 32 bit
Texture Filter: Trilinear

Result 97.2 fps (I'm running version 1.30, it seems that it has brought a slight drop from my usuall in 1.27, although this could be caused by any number of things).
I ran the tests four times thereafter for a total of 5 and the fps never deviated. This seems very strange given the fact that THG reports 1.30 to have wildly inaccurate results.

So is this a matter of my cpu? Simply put: NO. I ran some cpu scaling tests. At 1650Mhz (150x11) the score rose to 97.4fps a measly .2fps increase that can hardly be called a gain. I ran this test 5 times and no change.

But these are the lowest scores i have ever seen. By simply disabling lightmap the score jumps to 150fps and i have seen scores in the 160's-170's quite often.

Let me also say that the Quake III fps scales very well with the Kyro II clock settings. Let me mention here that the Kyro II's clock/mem speed is locked therfore i will only use one number when refferring to the overclock.

At 1466mhz(standard speed for AXP1700) MAX settings:
175Mhz(Base)- 97.2 fps
180Mhz - 100.0fps
185Mhz - 102.2fps

I ran the tests at each resolution 5 times and the results were within .1fps each time.
If you would not the Kyro II scales very linearly- Gaining about 2.5 fps each time. I did not raise the card above this point because it can't handle it very well. These cards are already pushed to the max.

Here are my computer specs (for your entertainment)-
AXP 1466 [/w PAL8045 heatsink :) , it has nothing to do with performance, but who the hell cares?]
Gigabyte 7Vtxe motherboard (which has been getting bad press, although it overclocks very well)
512MB DDR Ram (2x256MB OCZ 2400)
Hercules 3d Prophet 4500 (Kyro II)
Western Digital 40BB
Netgear FA311 (why did i even put that in there?)
Enermax 450Watt power supply (bored yet?)
CTX 1565 monitor (you get off on this don't you?)

BTW, My 3dmark 2000 score is 71xx (fluctuates)
3dmark 2001 35xx


:eek:  <font color=blue>I for one run Quake 3 on a P133(No MMX)</font color=blue>I have no affiliatioin w/ Intel

More about : kyro misrepresented

Anonymous
a b U Graphics card
December 21, 2001 3:49:00 AM

One question, did you set the 85mhz refresh rate the benchmarks came from?

It never really says what test setup the benchmarks for the KyroII came from, but looking back at the Hercules KyroII review (http://www4.tomshardware.com/graphic/01q2/010425/kyroII...) it looks like the reviews came from that system setup.

CPU: AMD Athlon 1.20 GHz, 133 MHz FSB DDR
Motherboard: ASUS A7M266 (AMD760 chipset) Bios: 1004 C Beta
Memory: Micron 2 x 128 MB DDR-SDRAM, PC2100, CL2.5
Hard drive: Seagate ST313021, 12 GB, 5400 U/min

Also windows settings themselves can make quite a difference (benchmarks used win98).
Anonymous
a b U Graphics card
December 21, 2001 2:27:10 PM

Hmmm, no i did not set the refresh rate to 85hz. But this should not have an impact, simply because the VSYNC is OFF. I remember there being issues with VSYNC with certain Kyro drivers, on which if you set the VSYNC to on it would actually be off and vice versa. This could very easily explain why THG got 84.9fps with 85HZ refresh rate.

I'm gonna email them and ask what drivers they were using.

As for the system setup. I don't think that makes a difference. 256MB ram is more than enough, and as for the difference in CPU speed, that is negligible. In the original post i ran my comp at 1466 and 1650 and the results changed by .2fps. I think the drop to 1200 would have a similar effect. I cannot simulate this however, because i cannot downclock my computer.

THX for the reply though.

:eek:  <font color=blue>I for one run Quake 3 on a P133(No MMX)</font color=blue>I have no affiliatioin w/ Intel
Related resources
Can't find your answer ? Ask !
December 21, 2001 2:38:47 PM

I can only offer opinion on this issue and I'm not trying to pass it off as fact.

Tom's reviews sometimes do suck and sometimes are inaccurate. IMHO Anandtech has the most comprehensive and accurate reviews.
Anonymous
a b U Graphics card
December 22, 2001 1:54:31 AM

Well, Anandtech has 100.3 FPS 1024x768 normal settings. I'm not sure how much max settings would lower the fps, but it seems in line at least with tom's.

I'm sure both sets of numbers used early drivers.

<P ID="edit"><FONT SIZE=-1><EM>Edited by Burns on 12/21/01 11:01 PM.</EM></FONT></P>
Anonymous
a b U Graphics card
December 22, 2001 3:18:34 PM

Seems strange, oh well. In any case, point- Kyro II does very well.

:eek:  <font color=blue>I for one run Quake 3 on a P133(No MMX)</font color=blue>I have no affiliatioin w/ Intel
December 22, 2001 3:39:15 PM

Kyro 2 uses a lot of CPU power, therefore, the increase may be normal.

What if your life moved.....2 inches to the left?
December 23, 2001 3:15:59 PM

Each 5 HZ in video refresh rate decreases performance by about 1.25 frames per second. That's with Vsync turned off. The video card has to spend more time doing video refreshes. This steals GPU clock cycles and slightly decreases time on 3D processing. Going from 70hz to 85hz would lower framerates by about 4 frames per second.

Update:
<b>Don't know what I was thinking! That's 1.25 fps second for every 15hz!!!</b>

<b>We are all beta testers!</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 12/23/01 12:47 PM.</EM></FONT></P>
December 23, 2001 3:27:27 PM

Yes! Go LCD's! To sync, they have refresh of 60Hz!

What if your life moved.....2 inches to the left?
December 23, 2001 3:36:58 PM

How are LCD panels in games these days?

I was just at a computer show yesterday and noticed how gorgeous the desktop display was on a $300 USD, noname display (15 inch). The computer doing the display didn't have any games.

How's the quality control. It would drive me crazy if I got flat panel which had couple pixels stuck on.

<b>We are all beta testers!</b>
Anonymous
a b U Graphics card
December 23, 2001 3:43:37 PM

Woohoo! An extra 6 FPS!
December 23, 2001 5:31:48 PM

Like flame said, Kyro2 is very CPU dependant. And various settings in BIOS and windows will also make an impact. All the reviewers' GF3 score are awful low compare to mine, because they tend to leave everything at default and I like to tweak my settings :) 
Anonymous
a b U Graphics card
December 26, 2001 2:44:42 AM

Sorry, if you haven't noticed yet i come and go, so sorry i didn't reply.

I got an email from the author saying that the test was run at 60hz (not 85?). So that blows my theory about VSYNC out the window.

As for CPU power? IMHO, and i can only assume: i ran tests at 1466 and 1650. There was a .2fps difference. I can only ASSUME that the drop to 1200 wouldn't have that much of an impact?

BTW, i do know about the issues with 3dMark 2001 render to texture and Kyro II, i retract my score.

THX for listening...

:eek:  <font color=blue>I for one run Quake 3 on a P133(No MMX)</font color=blue>I have no affiliatioin w/ Intel
!