Do you just inflate your benches then throttle down?

[ChrisRay]

Lately I have seen alot of stuff occuring in benchmarks (well not just lately as far back as I can remember)

People who set their computers to unrealistic levels. Then benchmark them. Then downclock them again just for a few more 3dmarks/Aquamarks, (this is notoriously bad in 3dmark/aquamark senses. Also setting LoDS super low.

I have a philosophy when it comes to benchmarking/pressing my own numbers. Never bench in a setting I wont play in. I have this bad habit of playing in a setting (Like 350 core on my 6800NU) for just a few days then downclocking it, (Why? Probably because I cant its perfectly stable and actually runs cooler) Currently I am testing myself to see if I can leave my system @ 350/900 for a week, Once I get used to it I leave it like that no problem.

In Any Event, I have a real grudge against people who bench there system in unrealistic enviroments. And game in other systems. Because at that point. The benchmark itself becomes pointless for comparing systems.

[GlowStick]

I disagree, i only overclocked my 9800 Pro just to see what the highist 3dmark03 score was i could attain. When i found that the best oc yeilded very little gain, i decided its not worth it. However on my cpu i have it overclocked all the time.

[ChrisRay]

But doesnt that just further make the benchmark more useless? I mean if the results are supposed to be comparative, And the enviroment is unrealistic, Why do it?

[nvnews-reader]

I think it's still useful you just have to make sure you are comparing apples to apples. The main reason I run benchmarks is to make sure my system is setup correctly and giving me the performance that my hardware is capable of doing. I'm not a big overclocking fan. I would rather keep the warrantees and reassurance that things aren't going to break prematurely then. The extra 10% performance or whatever I could get by OCing isn't worth it to me.

[GlowStick]

Originally Posted by ChrisRay

But doesnt that just further make the benchmark more useless? I mean if the results are supposed to be comparative, And the enviroment is unrealistic, Why do it?

No, i dont flash my 3dmark scores around on my sig, i mainly keep them for my own refrence of what my card can do. And actually i decided to keep my oc on my bfg card : )

[SH64]

Chris i agree with you .. thats exactly what i'm doing .. i always like a real-life scores .. not just to "super clock" the CPU / GPU / RAM & then take the score just to show off how powerfull is my system .

heck i even take my benchs in the worst conditions .. like with internet connection active & the room temp is too high (A/C off) .. so when i play a game i know how my system is going to handle it with & w/o overclocking.

[Kamel]

i wanted to just see how high it would go... was impressed, and since my cooling is so excessive i really didn't care to play normal games with my settings as they were when doing so (to be honest, i'm quite pissed off that it makes you "test" the settings first, because i feel i can overclock much higher). my mouse started messing up though, so i turned overclocking on automatic to see what would happen... it didn't fix it, but i haven't yet changed it back.

anyway, in short, yes, i will definitely play games at the settings i do on benchmarks, i try to make everything default when i bench so the results are as accurate as possible (if something is messing up, i want to fix it).

[ricercar]

"Inflate," what's that? Synthetic number generators like 3DMark and Aquamark are designed for quantity, not quality. All's fair in Onionmarks and war. There's no ambiguity in my world. Benchmarking is why God invented "Best performance" contol panel settings.

But benchmarking has no relevance to my daily gaming. "Best image quality" settings is my daily fare.

But even beyond that, my mere 5700 Ultra is in my main machine, while my 5800Us and 5900Us are currently in the closet. The convenience of a dual-DVI single-slot card overrides any "performance enhancement" from DVI+VGA dual-slot monstrosities.

[jAkUp]

Its more of a pissing game.. who can get the highest.. etc. Believe it or not.. I dont even play at the specs in my sig! Well.. I could, but I prefer everything to be ultra stable..

[ChrisRay]

Originally Posted by jAkUp

Its more of a pissing game.. who can get the highest.. etc. Believe it or not.. I dont even play at the specs in my sig! Well.. I could, but I prefer everything to be ultra stable..

Why am I not surprised you are the very essence of everything I hate in life!

Hehe j/k

Thats why I only keep my 6800NU @ 325/900. I havent seen a significant gain in the games I play to require anything else. The 900 Mhz memory was just a formality of restoring my 5900's bandwith, But I wanted 28.8 gigs of bandwith..

[OWA]

I'll do that but I don't ever use settings I wouldn't use in a game (i.e I don't use settings that aren't stable). I'll use my maximum overclock that is stable and artifact free for benchmarking but usually use a setting less than that for everyday use. In the real world the higher overclock just changes the framerate by a few fps so I don't see a need to keep the card at that level (even if it is stable). I do enjoy trying to get the highest score I can which is why I bother with the higher overclocks. I was playing around with this recently and was able to get a 1000+ plus higher in 3DMark03 by overclocking but of course, like I mentioned already, this only changed actual game benchmarks by a few fps. I was comparing a stock GT to 415/1160.