Printable View
My score seems a bit low for my settings doesn’t it (look sig)? I m’ getting 32sec in super pi 1M.
The card is a vivo modded to 16 pipes @ 1.7vgpu (artifact free @ 600/600).
I see people with a clawhammer at 2500-2600mhz getting 1 to 2k more with the same vidcard and clocks.
My dragon seems lower than most people here..
http://service.futuremark.com/compare?2k1=8161226
Any hits tips?
I can't get my score up :/
Here's a cpuz shot
http://home.planetinternet.be/~emeynen/CpuZ.jpg
Thx in advance!
That makes 2 of us im getting 28.4k dunno really whats wrong here...
http://service.futuremark.com/servle...&projectType=6
you got 30k thats more then enough
Have you fully tweaked the dfi memory timmings?
thats because your cpu is a newcastle, 3dmark lovvvvvvvvves 1mb cache, especially with x800xt cards. If you had a claw youd be a few thousand higher
I have a newcastle @ 2700 on air and get just over 30k(30100~) with stock xt pe clocks, if it takes you over 600/600 clocks and 2900mhz to get there, there is definately something wrong
Running the RAM on a divider is hurting your score. Try and get your mem. clocks up and do some runs 1:1.
Why don't I believe you :stick: . I ran 3DMark2001se with my NC@2.9Ghz 11x264, DDR528, and a stock XT PE.Quote:
Originally Posted by WildKard
And got this score: http://service.futuremark.com/compare?2k1=8167236
Quote:
Originally Posted by WildKard
"Why don't I believe you . I ran 3DMark2001se with my NC@2.9Ghz 11x264, DDR528, and a stock XT PE."
I believe him. First of all, your winxp just kills performance. Yes you may have 2.9GHz but his 2.7GHz and win2k is gonna be just as good. 2nd of all, he may be running 270x10 with bh5 or other low latency ram which would again kill your score. and thirdly gpu clocks do not matter except in nature and a stock x800xt pe should have 260fps in nature anyway. You have much tweaking to do
by the way, 512k newcastles really suck at 3dmark with a high end gpu, my friend eshbach has over 32k with only 2.56GHz and x800xt both air cooled and hes beating you even when he had his cpu at 2.4GHz he broke 30k, this is higher than your 28k at 2.9GHz
But I've got a Clawhammer score too you know :p: http://service.futuremark.com/compare?2k1=8049314
ThommY, how yo manage to use DX8.1 for DFI NF3 250Gb when the nvidia drivers etc need DX9 installed ???Quote:
Originally Posted by Geforce4ti4200
While they suck at 3d2k1 compared to Clawhammers, newcastles are okay, my 9800Pro R350 card has a decent 3d2k1 on the ORB given the drivers and DX9 used http://service.futuremark.com/compare?2k1=8179979
Umm, not necessarily. My 3000+ only goes up to 9x, so I have to use a divider to hit decent CPU speeds. ;) http://service.futuremark.com/compare?2k1=8186909 @ Thommy, tweak tweak tweak. :DQuote:
Originally Posted by Frank E. BoNeS
I was just thinking, do you have 1T enabled? because if youre running 2t that will absolutely KILL your performance, its worth like 35 FSB with equal timings
Also try running your tests in one of these 2 orders
LL DL CH DH LH N CL or
DL LL CH DH LH N CL
Some systems DL runs best first, sometimes it is LL, I run DL first
Cant wait for my dfi to get here. Here is mine with 520/560.http://service.futuremark.com/compare?2k1=8180818
My dragon low score is just to low. The rest seems ok. Wierd:/
find a score
here all cpu and fsb dependant test are below so can something be with your rams or system tweak
http://service.futuremark.com/compare?2k1=8187288
Quote:
Originally Posted by ThommY
weve been trying to tell you its the newcastle's 512k cache killing it. There is like 300MHz difference between that and a claw. Meaning a claw at 2.6GHz is gonna score just as high as your newcastle overall with it being ahead in drag low and car low probably. You could try a claw if you can afford one I guess
Hehe... GF4 your so funny :P
Win2k + the cache isnt gonna add a couple 1000 points...
Win2k adds like 2-300... maybe 500 with FULL tweaking... But vs a WinXP system FULL tweaked i doubt its even that...
And the card doesnt matter with clocks... Its a 100% CPU block in 3dmark01 ;) Maybe with like 5-6Ghz it counts...
The cache i would say is more like 150Mhz difference....
But hey 3dmark2k1 is dead... Its a crappy bench... Its no longer a 3Dmark... Its just a CPU mark now ;)
Might as well run superpi/pifast...
he has win2k and hits 30.7k at 2.9GHz, let me show you the scores from our own eshbach 32k at 2.56GHz so your wrong, the cache matters 300+ MHz. If I benched my cpu I would be lucky to even break 29k at 2.6GHz but a 2.4GHz claw can get me 30k. I dont even know if ill hit 500fps in drag low with this newcastle, "they" were right, the loss of cache really eats into your tests, especially drag and car low. also if 3dmark 2001 was so cpu limited, why am I below 21k yet? gpu does make a difference, just alot less than in 2003 and 2005.
http://service.futuremark.com/compare?2k1=8158853
32015 from eshbach! I compared his vs. Tommy's and my god eshbach is owning Tommy's score across the board except for car high where Tommy is 8fps ahead due to having 340MHz more cpu power but the 512k cache kills performance in every other bench, just car high cares the *least* for cache. With this I just saw, there is NO way I am getting a high end gpu without a 1mb cache cpu(or 512k if its 400MHz faster) meaning say when I get a 3.4GHz winchester or 3GHz san diago then itll be good for a faster gpu. My cpu would not only bottleneck a high end gpu in 3dmark, but in games as well. I probably have the power of a 2.3GHz clawhammer at best
Geforce4ti4200 guess your pretty good at this what do u think is wrong with my score? got a 3400 1mb cache@2500, x800xt-pe@ 590/600 Stock Cooled 512 memory kingston valueram - http://service.futuremark.com/compare?2k1=8167074 :(
No one has said much about the 1T/2T RAM timings? I thought that killed performance... Could that be it?
I think it's mostly a question of improper tweaking. Max hardware out, set detail settings in Advanced Display Properties to absolute lowest, close explorer.exe (and anything else you can), and enjoy... If you've done that, then I can't help anyone. :p:
Quote:
Originally Posted by Godmyster
Besides pointing the obvious(winxp) it looks like either your ram is async or the timings are bad. There is the possibility your winxp install is dirty and something in the background is snapping up cpu cycles. It is best to reserve another hdd or partion with win2k just for 3dmark like I am doing, I have a 3dmark hdd for just 3dmark and a few tweaks for it such as rivatuner. You should be getting 30k with winxp or 31k with win2k
Im trying to get a HDD for Win2000 :) ill keep in toutch when i get it.. danke :)
Ram is on 1T here. I think I’ll get a clawhammer some day.. But I’m sticking with the newcastle for a while anyway :)
Thx for the help and keep it cool!
you are welcome. maybe wait for the san diagos? they are 90nm with 1mb cache on s939 so were looking at 4GHz with dual channel and 1mb cache with phase change or over 3GHz on air cooling. I dont care what others say, ive seen Ricky and Pedro hit over 3.5GHz on old .13u technology, theres no reason 90nm san diagos wont be seeing 4GHz then
Yeah, ... he's a regular legend in his own mind.Quote:
Geforce4ti4200 guess your pretty good at this what do u think is wrong with my score?