Low Benchmark in 3DMark '05

[Tatty_Two]

nice.with what voltage you have them?did you try raising the vdimm and try CL2?
i`m starting to think that my UCCC are not so good...

I drop that down to 2T for a little test and try goin 3-3-7, may not work with that overclock unless its not an overclock but very fast ram?

I am gonna leave you all in peace and not visit this thread anymore cause memory bandwidth is even strtin to bore me!!!! Cheers guys happy tweakin!

 

[Satchmo]

its all stable at T1 2.5,3,3,6

i tried cas2 but it wouldnt boot, im happy with what i have, @ 89% efficiancy and ~6400 mbps both ways.

my CPU score in 3dmark 05 went from 4400 to 4800 =) too bad it doesnt effect socre in 05.

my 06 went to 2965 from 2908, CPU score does affect over 3dmark score. so thats cool.

 

[Tatty_Two]

Well here is a story! You all know by now that I have cheap corsir value ram and I know I promised not to add another RAM message to this thread but I'm pretty damn pleased with meself! I have managed to clock my cheapo ram not only to 425Mhz but look at these for timings:

2-3-3-6 @ 1T!!!!! am I impressed or what, bandwidth now at almost 7000 running at 96% efficiency in Sandra, eat your heart out you exspensive Ram buyers!!!! must have got a really good batch is all I can say, was going to buy another matched pair to bring me up to 2 Gig but I am going to leave it for a while now as they surely wont give me these timings.

CPU-Z memory thumb to confirm timings attached although it dont show 1T.

 

[giorgos th.]

how much did the ram cost you?
btw nice results.try 2-3-2-6 1T with some better values in A64 tweaker.it will do better i believe.

 

[Tatty_Two]

how much did the ram cost you?
btw nice results.try 2-3-2-6 1T with some better values in A64 tweaker.it will do better i believe.

Thanks will give it a try, it cost me £45 for the gig, thats about $78, its also raised my 3D Mark 03 score a little, not tested it on any others yet, thumb attached.

 

[Satchmo]

http://www.newegg.com/Product/Product.asp?Item=N82E16820161679

I have two of these, in dual channel mode, 7200 bandwidth 2.5,3,3,6 T1 DDR450 @ 2.65v.

Expensive ram buyers tell me again how you spend $200 for a 2gb dual channel kit PLEASE, i love a good laugh.

I cant run sandra for some reason...it just wont start so i cant say what the efficiancy is.

 

[Tatty_Two]

http://www.newegg.com/Product/Product.asp?Item=N82E16820161679

I have two of these, in dual channel mode, 7200 bandwidth 2.5,3,3,6 T1 DDR450 @ 2.65v.

Expensive ram buyers tell me again how you spend $200 for a 2gb dual channel kit PLEASE, i love a good laugh.

I cant run sandra for some reason...it just wont start so i cant say what the efficiancy is.

Who spends $200 on a 2 Gig Dual channel kit< they are even cheaper than that in my local store, well UK£ equivilent I'm sure you could get them down to CL2 with some "fine tuning"

 

[giorgos th.]

i bought my Transcend UCCC for $90 and i`m very happy with them.rock stable at 278 mhz 3-4-4-8 1T with 2.7V.i`ve got ~7800 bandwidth with only 80% efficiency but who gives a sh**....
it was a boost from my old Geil ultra platinum..

 

[Tatty_Two]

i bought my Transcend UCCC for $90 and i`m very happy with them.rock stable at 278 mhz 3-4-4-8 1T with 2.7V.i`ve got ~7800 bandwidth with only 80% efficiency but who gives a sh**....
it was a boost from my old Geil ultra platinum..

Just goes to show that "Brand" and price does not always make it the best!

 

[MD]

Hi, just wanted to put up my 3d mark 05 scores with my card, it still has a bit left i dont want to push it much harder after i laid out 500 bucks a while back (the wife might banish me to my truck to sleep lol). It's a built by ati all in wonder x-800-xt, not running to crazy of clocks as you will see, as it's oly got the factory cooling on it, 1500Mb kingston value ram, athalon 64 3000+ running at 2.4 Ghz. Anyways i thought it wansnt too bad of a score so enough rambling here's the link ..... take care ..


http://service.futuremark.com/compare?3dm05=1447396

 

[trog100]

memory bandwidth makes very little difference in 2005 scores with amd 64 chips.. neither does cpu speeds.. its pretty much all down to the grfx card.. this also translates to real life gaming.. the grfx card rules and is the be all and end all..

trog

 

[Tatty_Two]

memory bandwidth makes very little difference in 2005 scores with amd 64 chips.. neither does cpu speeds.. its pretty much all down to the grfx card.. this also translates to real life gaming.. the grfx card rules and is the be all and end all..

trog

In the main I agree but I got an extra 170 in 3D Mark 05 and about 350 in 3D Mark 03 over the course of some memory tweaks, all other system/software settings were the same, the start product on memory was 400Mhz 2.5-3-3-8 2T and I gained the points when I finished up at 425Mhz 2-3-3-6 @1T, speaks for itself really, as you said, the biggest gain was in 2003 but there was gain in 2005 albeit only about 2-3% but for a free gain that aint bad in my book! especially on value RAM.

 

[trog100]

yes a free gain is worth having.. but mostly it aint free.. most folks spend big money on expensive memory to get it.. then it aint worth having.. but even if it is worth having in a real gaming world low frame rate problem situation it wont help in the slightest.. back to the grfx card being the king thingy.. he he

i dont see the gains from memory tweaking in 2005 u see.. i see far less.. like about 50 point difference between memory running at 320 or 420.. the amd chip seems to have rewritten the rules as far as memory bandwidth goes..

i am a firm believer in "value" ram.. especially if it runs cas 2.. most of it dosnt.. the problem being that all sorts of system instabilities can creep in by trying to make it..

i spose the point i am trying to make is that with the amd 64 bit system all the crap published about which super memory u should buy is just that.. crap.. and in essence a marketing con and waste of money that would be better spend on a faster grfx card..

however i aint against getting something for nothing.. just against folks spending their money in the wrong places.. expensive memory being the worse case.. expensive mobos and cpus come next... he he he

in todays gaming world the grfx card really is the king.. and in theworld of 3Dmark 2005 it fills the same role.. my 3Dmark scores.. with memory at cas 3 for stability..

####

amd 3700+ cpu at 2.9.. x1900xtx card at 690/800.. nanya memory 420 at cas-3..

3Dmark 2000 = 31237

3Dmark 2001 = 33202

3Dmark 2003 = 19254

3Dmark 2005 = 11934

3Dmark 2006 = 5385

trog

 

[MD]

I'm running the a-64 and had borrowed some of the old man's nice corsair xms 3200
PN: (twinxp1024-3200c2) with advertised 2-3-3-6, and didnt really get much of a change
and found my kingston value ram to be really good, it even runs nice at cas 2 (single 512)
in at a time with no trouble and it seems I get great performance to the buck ratio,for about 50 bucks canadian per chip. I dont know what my pops spent on his ram but i'm sure it was
alot more than he wants to admit after trying mine

On the other side i did get a little better 3d-05 score running 3x 512 sticks of value ram than two, but nothing i could really tell in game too much.

 

[trog100]

i think it comes down to all reviewers useing top cpus and top end memory for their gaming reviews.. it creates the impression u need such things when in actual fact u dont.. intel system are held back by memory bandwidth.. amd 64 systems dont seem to be..

i can run my system at cpu 2.4 gig memory at about 175.. or cpu at 2.9 gig memory at about 220 and in a game i cant tell the difference.. the difference in a 2005 score isnt much either.. grfx card tweaks/overclocks are noticable.. nothing else is thow.. with the proviso u have enough system memory.. if u dont have enough things kinda slow down dramatically.. he he

the bottom line is any default amd 64 cpu and two gigs of value memory teamed with an expensive grfx card will game nicely.. the money saved on the cpu and memory should be pumped into a better grfx card..

trog