Post your 3dmark06 HD 3K Series Results

[mikek75]

LOL, cheers. I might hold you to that!

 

[Dr. Spankenstein]

Hey mikek75,

Did you ever get your refresh rate sorted?

 

[trog100]

Just so we can put this to rest, what form of proof do you require?

Link for comparison.

Look at it yourself, although it is unfortunate that FM can't read these cards speed...


p.s. take a screenshot of our comparison. I'm sure someone will be able to decypher what's going on.

i dont want "proof".. stop being so bloody defensive.. way back in this thread someone ask me why my score was higher than theirs.. i gave them a simple explanation.. how about u giving me the same consideration..

i only have the one cpu freak.. E6750.. not sure how its got to be called a 6700..

trog

 

[Dr. Spankenstein]

Not being defensive, chum. Just looking for the snide comments to end.

Please humor me and provide a screenie of our comparo.

 

[trog100]

humour me and tell me how u get such a good score.. i am genuinely curious..

trog

 

[Dr. Spankenstein]

Yhgpm.


BTW, the single reason to benchmark is for competition. If we gave ALL our secrets* away, we'd all score the same! We could save ourselves many hours of watching dancing sprites and flying boats and just plug our hardware specs into a grid: "Hey, my system can get 15,000 3DM06 marks! WHOOPEE!"


*secrets, not "cheats" as you're so fond of referring to them!

 

[trog100]

Yhgpm.


BTW, the single reason to benchmark is for competition. If we gave ALL our secrets* away, we'd all score the same! We could save ourselves many hours of watching dancing sprites and flying boats and just plug our hardware specs into a grid: "Hey, my system can get 15,000 3DM06 marks! WHOOPEE!"


*secrets, not "cheats" as you're so fond of referring to them!

for the good of all.. we might learn something usefull..

http://service.futuremark.com/compare?3dm06=4371289

http://service.futuremark.com/orb/projectdetails.jsp?projectType=14&projectId=4577912


trog

 

[Random Murderer]

for the good of all.. we might learn something usefull..

http://service.futuremark.com/compare?3dm06=4371289

http://service.futuremark.com/orb/projectdetails.jsp?projectType=14&projectId=4577912


trog

the 2nd link doesn't work.
and i don't get it. i've looked at the result screenie on page 3 and his orb link 20 times, but i don't see this "secret" he speaks of... everything looks normal...

 

[Dr. Spankenstein]

the 2nd link doesn't work.
and i don't get it. i've looked at the result screenie on page 3 and his orb link 20 times, but i don't see this "secret" he speaks of... everything looks normal...

...my point exactly!

 

[Dr. Spankenstein]

@ trog100

Must look like this. Copy this link that appears after you click "UPDATE".

 

[trog100]

...my point exactly!

interesting thow.. the second probably needs logging into.. not important thow.. all the info is in the first one..

i dont see a secret.. i still see a wee mystery thow.. each grafix test consistently scores a little higher for no valid reason.. the cpu test scores a little lower for a very valid reason.. its going slower..

we have two identical cards mine is clocked higher.. mine should score more but for some reason it dosnt..

i dont think hardrive speeds affect the tests.. memory speed might but not by much.. running after a fresh reboot is normal.. 2006 can give odd low scores for no apparent reason sometimes.. a reboot always cures this..

so i have to admit to being completely baffled as to why a slower clocked card and slower cpu is churning out more frame rates..

if your system is going faster than someone else's its easy to say why something scores higher.. when its going slower it gets difficult... he he he

how is your fur bench score DR.. ???

still as curious as i was.. but thanks for cooperating..

trog

ps.. i only have the freeware 2006.. it dont kinda let me do much at all.. getting very late over here off to bed.. catch u tomorrow..

 

[yogurt_21]

does that mean my 1680 x 1050 widescreen is running more pixels than it should be then.. or am i loseing the plot somewhat here..

trog

ummm if your 3d06 defaults to that it would be the first time I've seen a screen do that. ( as usually its that the monitors default driver isn't capable of the verticle pixel amount on widescreens and 1080 is obviously higher than 1024 so that shouldn't happen). but yes you'd be running 1,782,000 26% more pixels slowing your performance and lowering your score.

@ dr. spankenstein your 2.0 looks normal to me
http://service.futuremark.com/compare?3dm06=4371289

check it vs my 2900 at 940 core and see what i mean

http://service.futuremark.com/compare?3dm06=4479694
while my quad give me an obvious score advantage it shouldn't affect the game scores at all, infact with your extra 100MHZ your cpu should be giving your game tests a bit more performance than mine. but the stream processor architecture has always favored sm3.0 over 2.0. it's just the way the drivers are optimised.

 

[unsmart]

pencil mod Vcore to 1.287v I'll probably up it to 1.3 in a few days. Whats a good under $30 HSF for this?
My CPU is killing my score

 

[Dr. Spankenstein]

@ yogurt 21,

Thanks for dropping the link on us. And your perspective.

A most respectable single card score! You crush my CPU score by 2100 points!

 

[yogurt_21]

lol yeah but the quad counts for like 70% of that and the other goes to my watercooled bios vmoded card lol. so in short mines good for checking 2.0 to 3.0 but I really "cheat" when it comes to voltages and cooling. lol

 

[Dr. Spankenstein]

i only have the freeware version.. so i could only do the defaults..

trog

Just go back in and select the "Published" box, click "UPDATE", copy the link that appears. Post.

dont see a secret.. i still see a wee mystery thow.. each grafix test consistently scores a little higher for no valid reason.. the cpu test scores a little lower for a very valid reason.. its going slower..

Again, give US some numbers to compare and the specs you wre running the test at.

we have two identical cards mine is clocked higher.. mine should score more but for some reason it dosnt..

It's been proven that some memory speeds kick in a higher latency. Maybe you've crossed over that point.

i dont think hardrive speeds affect the tests.. memory speed might but not by much.. running after a fresh reboot is normal.. 2006 can give odd low scores for no apparent reason sometimes.. a reboot always cures this..

There has to be some difference though. Try running 3DM from your floppy drive! (Shouldn't affect your score too much!)

so i have to admit to being completely baffled as to why a slower clocked card and slower cpu is churning out more frame rates..

Could be lots of things. I asked these same things to Random Murderer: What strap are you in? What FSB are you running at? What's your RAM running at?

if your system is going faster than someone else's its easy to say why something scores higher.. when its going slower it gets difficult... he he he

How would we know! You're about the only one here who hasn't submitted a screenshot with the pertinent information. CPU-Z, GPU-Z, CCC or Riva. Give us something to compare.

how is your fur bench score DR.. ???

Don't know. I'll try and post up.

 

[Dr. Spankenstein]

lol yeah but the quad counts for like 70% of that and the other goes to my watercooled bios vmoded card lol. so in short mines good for checking 2.0 to 3.0 but I really "cheat" when it comes to voltages and cooling. lol

Am I a "cheater" because I pencil-modded my card?!?

 

[The_Real_DeaL31]

i just hit 10k in 3dmarks06 with a hd3850 woot woot hehe

 

[mikek75]

Hey mikek75,

Did you ever get your refresh rate sorted?

Nope, still obstinately stuck at 60HZ! I have high hopes for Cat8.1 though, lol

 

[Tatty_Two]

Also, just to comment on one earlier point, in 3D Mark 2006 memory places a reasonably significant (ish) part, an extra 100mhz on the memory speed can give you MUCH more than 100 points, those of you with decent memory dividers in BIOS might like to try this as a little example:

Set your CPU speed at say 3.2Gig and run your mem really low on the lowest divider, then run the test, then set your CPU at 3gig and run the memory on higher (at least 200mhz higher to compensate for the 200mhz drop in CPU speed) but both memory speeds with the same latency, then compare your scores, this only really works on a dual core because on a quad, the CPU score takes a more signifcant drop with the reduction in cpu speed.

I can only speak for myself but my first run with my 8800GTS 512mb got me 17,022 points, to be safe, the memory divider I had on was running my memory at 1120mhz at 4-4-4-10, the next day, whilst playing around I set the card and the CPU to exactly the same but upped my VMem voltage and ran my memory at 1380mhz on 5-5-5-15 (2.45V ), I scored 17,211, almost a 200 point increase, had I run the memory at 1120mhz on 5-5-5-15 rather than the 4-4-4-10 I think you would have seen a 400 point improvement.

 

[trog100]

How would we know! You're about the only one here who hasn't submitted a screenshot with the pertinent information. CPU-Z, GPU-Z, CCC or Riva. Give us something to compare.

my first post didnt have anything in except the 2006 score.. the other two show my card and cpu speeds thow.. CCC.. CPU-Z.. and the 2006 score at default settings..

u have enough info to work with.. just to refesh

my memory is running cas 5 at 950 mhz 1 to 1..

my fsb is at 1900.. cpu at 3.8 gig..

my card at 895 core 1275 memory..

i dont know how u are running your memory but pretty much everything on my system is running higher clocks than on your system..

things vary a bit but clock speeds are what its all about.. especially card clock speeds.. my extra card speeds should give me roughly 100 point over your score.. my extra cpu speed should me perhaps 50 points over your score..

i should be getting 150 points more than u.. but u are getting 200 points more than me.. so we have a discrepancy of 350 points..

so where does that come from.. its possibly from memory speed and latency variations.. i dont know your memory speeds or timings.. ?

to get my score up to where yours is i would in essence have to run my card at about 915 mhz.. pretty high clocks..

soooo... both these scores are reasonable scores.. yours is better than mine.. i am simply interesting in finding out why at lower cpu and card speeds u score more than i do..

pure research.. nothing more..

trog

 

[Dr. Spankenstein]

my first post didnt have anything in except the 2006 score.. the other two show my card and cpu speeds thow.. CCC.. CPU-Z.. and the 2006 score at default settings..

I see that now, my bad! But now, going by your screenshot your card shows to be running 895/1275. Mine is at 877/1287. So yes, your card's core is slightly faster. For the record, the card's core goes up in 13.5MHz increments, so you're only getting 891 (Check with RivaTuner!) I'll conceed the 13MHz core.
Your card's mem is running slower, which we all know contributes to the bandwidth that the card can transfer information. I've seen quite large gains/losses from the same core speed with different mem speeds. So your theory loses a little credence there.

And 50MHz CPU speed isn't such the huge difference that you make it out to be.

I've had variations of 200 points between different runs for no apparent reason.

I think both our scores are where they should be. In fact, your score and mikek75 seem to be scaling quite proportionately, seeing as you both have the same CPU.

There is ALOT to be said about having, what can only be described as, a "balanced" system. It's not always a matter of jacking your FSB to the max or cranking your memory as tight as a drum.

There are others that will support me in this statement since I'm fairly new to all this myself. I don't claim to have all the answers, but to get your knickers in a twist over (what in the end analysis) amounts to 200 3DMark points seems a little zealotous.

Just sayin'...

 

[Lillebror]

Everything at standard - Havent overclocked Cpu yet.. hoping for 11k when its overclocked!

 

[InPaniC]

Hi,

i love this HD3870, much more power than 2 2900XT´s and very cold on wakü:

(Vista costs so damn CPU points...) And its CF of course

 

[trog100]

I see that now, my bad! But now, going by your screenshot your card shows to be running 895/1275. Mine is at 877/1287. So yes, your card's core is slightly faster. For the record, the card's core goes up in 13.5MHz increments, so you're only getting 891 (Check with RivaTuner!) I'll conceed the 13MHz core.
Your card's mem is running slower, which we all know contributes to the bandwidth that the card can transfer information. I've seen quite large gains/losses from the same core speed with different mem speeds. So your theory loses a little credence there.

And 50MHz CPU speed isn't such the huge difference that you make it out to be.

I've had variations of 200 points between different runs for no apparent reason.

I think both our scores are where they should be. In fact, your score and mikek75 seem to be scaling quite proportionately, seeing as you both have the same CPU.

There is ALOT to be said about having, what can only be described as, a "balanced" system. It's not always a matter of jacking your FSB to the max or cranking your memory as tight as a drum.

There are others that will support me in this statement since I'm fairly new to all this myself. I don't claim to have all the answers, but to get your knickers in a twist over (what in the end analysis) amounts to 200 3DMark points seems a little zealotous.

Just sayin'...

my knickers aint in a twist.. my score does scale very well with mikes.. yours dosnt.. its factors like this that make me look extra close at your score..

the cpu speed difference isnt 50 mhz.. its 100mtz.. u are correct my gpu speed according to riva tuner and gpu-z is 891.. still significantly higher than yours.. the very small memory speed difference is totally insignificant which is why i dont mention it..

my system is well balanced more so than yours.. basically u dont have the faintest idea do u.. u are clutching at straws trying to find a reason for your higher than it should be relative to mine score.. he he he

i aint new to this.. my 2006 score is consistent it doesnt vary by 200 points at random..

my system should score higher than yours.. it dosnt.. why.. neither of us knows..

it seems we aint gonna find out either..

the next time someone asks u why your scores are higher than he thinks they should be.. be honest.. just say.. they are high arnt they.. as to why i have no idea.. he he



trog