d3v0 - 16 Graphics Cards Tested: New/Old, stock/overclocked/vmodded - 9 Benchmarks - Page 11 - Overclock.net

d3v0 · 02-21-09

Haha thanks, I was afraid I would do that. I finished this at 4am

Ill fix it tonight.

XFreeRollerX · 02-21-09

just figured I would point it out

SLeeZeY · 02-23-09

Quote:

Originally Posted by d3v0

I would say that 1680x1050, High, 2xAA, dx10 mode is high for crysis, yes.

If you look, I posted my settings for each benchmark. All graphics cards were tested unerringly with the exact same settings to give the best side by side comparison.

Haha ok, missed that, but benchmarking a £60 card on those settings against HxC gamer cards is less of a benchmark and more of a slaughter.

Surely a 'reasonable' bunch of settings should of been made for the smaller cards.

StormX2 · 02-23-09

wow this is some excellent work!

Gives me a Good idea of what my current card can and cannot do.

As well as telling me whether or not i want to grab a GTX 260 or wait for something newer - Tough decisions lol Hoping that this i7 build I am waiting on will unlock enough of my GPU to make me happy for a while.

You knwo what I Absolutely wish I could see, is performance of your machine, with GTX 260, vs 8800 GTX on warhammer online with 200 people in RvR at 1680x1050 resolution =)~

THAT would make me super happy

d3v0 · 02-23-09

I am a WoWer not a WARman

d3v0 · 02-23-09

Quote:

Originally Posted by SLeeZeY

Haha ok, missed that, but benchmarking a £60 card on those settings against HxC gamer cards is less of a benchmark and more of a slaughter.

Surely a 'reasonable' bunch of settings should of been made for the smaller cards.

I am going to add excellent questions like this to my FAQ/Discussion section soon.

The reason I tested all cards equally under harsh conditions is due to the methodology stated in the original post. I wish to play games at their maximum in order to get full immersion. Jagged edges due to a lack of Anti-Aliasing or muddy low quality textures really grate on my nerves and I am sure that this sentiment is one shared by many gamers.

Thus, I decided to pit each card against each other in full-bore benchmarks. In compensation for the little guys, I made sure to overclock each one to their absolute stable limit. The only exception to this rule is the GTX 260 Core216, which I only cranked up to Evga's "FTW" clocks, which are extreme enough.

The purpose of this also was to be able to give an accurate comparison. For example, some people would say "lulul GTX 260 is 10x faster than an 8600gt!" when the reality in gaming is that it's only 4x faster. And in synthetic benchmarks, even less so (save for 3DMarkVantage GPU score!).

Also remember that the entire purpose of this spreadsheet/benchmark comparison was to give potential buyers a snapshot of the real performance gains they would see from their GPU upgrades. That is another reason why I overclocked my Q6600 to a reasonable 3.6ghz on a cheap board, and used DDR2 ram set only to 1000mhz. This was in order to provide the least possible confounding variables and most mainstream comparison possible.

Anyone with great questions like this, never be afraid to ask them.

I also fixed the Crysis Bottom 16 picture, the 9600GT OC is now dark green instead of dark red

SLeeZeY · 02-23-09

Quote:

Originally Posted by d3v0

I am going to add excellent questions like this to my FAQ/Discussion section soon.

The reason I tested all cards equally under harsh conditions is due to the methodology stated in the original post. I wish to play games at their maximum in order to get full immersion. Jagged edges due to a lack of Anti-Aliasing or muddy low quality textures really grate on my nerves and I am sure that this sentiment is one shared by many gamers.

Thus, I decided to pit each card against each other in full-bore benchmarks. In compensation for the little guys, I made sure to overclock each one to their absolute stable limit. The only exception to this rule is the GTX 260 Core216, which I only cranked up to Evga's "FTW" clocks, which are extreme enough.

A cool, I'm all for thrashing little cards is all, but the Visiontek HD4670 1GB - OC @ 825c/1000m is pretty dire lol, I think you should get the Novatech Palit version, I went for it coz it was the cheapest I could find, but it seems to be a little monster stock 700/1005 to 901/1200 stable (aslong as you can keep the temp down).

Again the same as my old EAH3650, the core will clock like nuts if you lay off the mem clock. If the core is untouched the mem will go up a lot more, which I presume is purely a voltage draw limit and altering the cards BIOS to raise the limit would let them both climb higher (but I'm too poor to be melting things for fun).

edit: I just added a link, seems to have come down another few quid, so I might be buying my second card for crossfire sooner than I thought

d3v0 · 02-23-09

Haha I had to buy what I could find. I couldnt find the reference version locally, same with the 9800GTX+ 1gb. Both cards I would have preferred to test reference models.

almighty15 · 02-23-09

Single stock clocked 8800GT 1gb with F.E.A.R benchmark..

1280x1024, 4xAA, 16xAF soft shadows on and everything else maxxed out

Min : 47fps
Avg : 99fps
Max : 248fps

1680x1050 is ~25% more pixels...so simulated 1680x1050 results would be

Min : 35fps
Avg : 75fps
Max : 186fps

Bit pointless but i did my math practice for the day!!!

d3v0 · 02-23-09

Haha umm well I'm not sure it works the same way as pure arithmetic but sure why not

Looks like your results fit fine

		Overclock.net - Overclocking.net > Graphics Cards > Graphics Cards - General
d3v0 - 16 Graphics Cards Tested: New/Old, stock/overclocked/vmodded - 9 Benchmarks

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)