Originally Posted by Kipman725 a return to meaningfull benchmarking :O
Agreed.
Whilst the 'highest playable settings' way of doing things was good for letting you know how your rig might be expected to handle certain games, or how a potential new purchase might be expected to perform, it doesn't beat the good old 'straight comparison'.
Us geeks then get to see how our FPS are affected in several games, several resolutions, various levels of AA & AF, and as far as I'm concerned, make a more informed decision.
It allows a much easier way of comparing hardware directly to see if that shiny new upgrade is really going to make that much difference. Rather than just saying "it allows you to play Quake 4 with 4xAA!"
That's not taking anything away from the previous method of benchmarking. I think it was 'ballsy', given your readership and probably took as much, if not more work to compile than this one.
I just think that the people who read a hardware article on bit-tech are big enough and ugly enough to make an informed decision themselves, once provided with some hard facts and figures.
Mr. Smalley, I salute you ;)
Oh, and I think I'll stick with my 8800GTS 640MB but the 320MB looks like a great deal.
EDIT: One thing I was curious about is that your screen grab for COH doesn't suggest nearly the highest graphical settings (though I accept this may be a library pic). I run the game with all settings maxed out on my rig except for 'effects fidelity' and 'effects density'. I currently have these set to 'high' but setting them to 'Ultra' brings my lil' beast to it's knees, with shockingly low frame rates, which render the game completely unplayable. Any thoughts?
hmmm I always liked the highest playable settings feature myself but eitherway, good stuff as always. This little bugger looks like a HTPC solution for some mild gaming....or generals on a 50" lcd ;)
Originally Posted by Bit-Tech Review BFGs GeForce 8800 GTS OC 320MB is not going to win all of the benchmarks with its fairly middle of the road clock speed increases, but BFGTech does appear to have hit a price point right on the top of the head. The card is available for purchase at just under £195 including VAT thats about £70 cheaper than the 640MB version of the same card with the same clock speeds.
Meh. Needs to be a little better. Looks like I'll be getting the 640mb or the GTX......Seems like my 7900GTX can pull those numbers albiet without DX10 support
Looks like a very nice card to me. A great spot at £190 for now.
I too, do like the return to simple numeric benches. The Best Playable thing was pretty useful in some respects, and if that could be partially maintained or maintained for some games I reckon it'd be cool. Not sure how though since it seems like an all or nothing method. Regardless, it was far easier to compare the 640 with the 320 using the method used in this review imo.
One thing though tim, can we get some Oblivion tests please, preferably with some super-hi res texture mod packs? I'd be very interesting to see how the 320 performs in Oblivion which I have a feeling might chew through RAM with hi-res packs installed.
good review, would consider it, as i may be upgrading my card from a meassely ati x300 :(, but would conisder the 640MB one just to get a few years out of it. or might just wait till a direct x10 card cums out, the r600 should be good!
Originally Posted by mctigger good review, would consider it, as i may be upgrading my card from a meassely ati x300 :(, but would conisder the 640MB one just to get a few years out of it. or might just wait till a direct x10 card cums out, the r600 should be good!
That's going to be an impressive performance jump!
No 1280x1024 benches then?
I understand that the market is moving to larger 1600x1200 panels but still. I'm sure the larger percentage here is still doing 1280x1024.
I'll stick with my 640 MB GTS too. If the 320 had been released earlier... I'd still have picked the 640. ;)
And looking at the overclocking potential those cards have (and the E6600) I think it's hard to get something better for the same money if you're actually willing to overclock a bit.
// edit: What I missed was the overclocking part in the review though. Did I over-read it? And I like the 'apples to apples' approach better than 'best playable settings' too. :)
Confused with the review, you are saying that if you play at 1280x1024 there is no use getting a GTX like I will be doing pretty soon? Or have I read wrong.
I havent upgraded for 3 years, and in March I will start ordering my stuff, which is going to be £1800 for the machine only. Which means I really want the thing to last as long as it can. Is the GTX worth the £180 more than this? IMO I think it is due to it being more future proof for games i.e. Crysis. Which I will def be playing.
As far as the benchmark style goes, [H] seems to be sticking with the highest-playable-settings. Personally, I wouldn't buy anything without at least two different reviews anyway, and checking both gives two different types of insight!
In their review, their card got slaughtered under Quake 4, where as here it held up under fire just as well as the others. Difference in settings? The overall trend seems to suggest Anandtech's results are awry; the ones here and at [H] suggest a slight disadvantage at best at 16x12, not a slaughter.
Great review though -- compared several major brands!
i miss the old testing method already. canned benchmarks don't reveal anything about the kind of toll features like ssaa or adaptive aa have on games. if you're paying upwards of 300 USD on a video card, wouldn't you at least try to turn on as many switches as possible. enable the more sophisticated aa features and tell me whether a x1950xtx is anywhere near a 8800gts 320. i mean, here in the states the lowest price for a sapphire x1950xtx from a reputable etailer is 380 USD including tax and shipping, the evga 8800gts 320 is 312 USD. canned benchmarks suggest that these two products are somehow on equal footing.
Originally Posted by MajorGN Confused with the review, you are saying that if you play at 1280x1024 there is no use getting a GTX like I will be doing pretty soon? Or have I read wrong.
I havent upgraded for 3 years, and in March I will start ordering my stuff, which is going to be £1800 for the machine only. Which means I really want the thing to last as long as it can. Is the GTX worth the £180 more than this? IMO I think it is due to it being more future proof for games i.e. Crysis. Which I will def be playing.
As far as I understood all the reviews of G80 it's not totally pointless to get a GTX to play @ 1280x1024. You'll get everything maxed out + AA + AF at awesome frame rates. Owning a GTS myself I have to say it's way enough at the moment for EVERY game I tested @ 1280 but if you want to last your coming PC for another 3 years go for a GTX as the next next-gen games (the ones after Unreal Engine 3.0) will definitly stress your GTX a lot. And within 3 years wouldn't it be realistic to say you might own a 22"-24" monitor? If you have the money to go for a GTX, do it damnit! :)
[offtopic]
£1800? That's about 2,687.89! What do you plan buying? Core 2 Quad, GTX SLI? I paid roughly 1,300 for my PC (see sig)...
[/offtopic]
Comments 1 to 25 of 35
Agreed.
Whilst the 'highest playable settings' way of doing things was good for letting you know how your rig might be expected to handle certain games, or how a potential new purchase might be expected to perform, it doesn't beat the good old 'straight comparison'.
Us geeks then get to see how our FPS are affected in several games, several resolutions, various levels of AA & AF, and as far as I'm concerned, make a more informed decision.
It allows a much easier way of comparing hardware directly to see if that shiny new upgrade is really going to make that much difference. Rather than just saying "it allows you to play Quake 4 with 4xAA!"
That's not taking anything away from the previous method of benchmarking. I think it was 'ballsy', given your readership and probably took as much, if not more work to compile than this one.
I just think that the people who read a hardware article on bit-tech are big enough and ugly enough to make an informed decision themselves, once provided with some hard facts and figures.
Mr. Smalley, I salute you ;)
Oh, and I think I'll stick with my 8800GTS 640MB but the 320MB looks like a great deal.
EDIT: One thing I was curious about is that your screen grab for COH doesn't suggest nearly the highest graphical settings (though I accept this may be a library pic). I run the game with all settings maxed out on my rig except for 'effects fidelity' and 'effects density'. I currently have these set to 'high' but setting them to 'Ultra' brings my lil' beast to it's knees, with shockingly low frame rates, which render the game completely unplayable. Any thoughts?
**cough, £45, cough**
I too, do like the return to simple numeric benches. The Best Playable thing was pretty useful in some respects, and if that could be partially maintained or maintained for some games I reckon it'd be cool. Not sure how though since it seems like an all or nothing method. Regardless, it was far easier to compare the 640 with the 320 using the method used in this review imo.
One thing though tim, can we get some Oblivion tests please, preferably with some super-hi res texture mod packs? I'd be very interesting to see how the 320 performs in Oblivion which I have a feeling might chew through RAM with hi-res packs installed.
Overall, great review :)
When scan have the Palit 640MB 8800GTS' in stock they normally put them on today only for £240.
Sam
extra year or two out of the card.
Great review.
Yeah, I realised later on.
That's going to be an impressive performance jump!
I understand that the market is moving to larger 1600x1200 panels but still. I'm sure the larger percentage here is still doing 1280x1024.
And looking at the overclocking potential those cards have (and the E6600) I think it's hard to get something better for the same money if you're actually willing to overclock a bit.
// edit: What I missed was the overclocking part in the review though. Did I over-read it? And I like the 'apples to apples' approach better than 'best playable settings' too. :)
I havent upgraded for 3 years, and in March I will start ordering my stuff, which is going to be £1800 for the machine only. Which means I really want the thing to last as long as it can. Is the GTX worth the £180 more than this? IMO I think it is due to it being more future proof for games i.e. Crysis. Which I will def be playing.
But one thing I don't follow. The benchies here don't match at all to Anandtechs; http://www.anandtech.com/video/showdoc.aspx?i=2926&p=8
In their review, their card got slaughtered under Quake 4, where as here it held up under fire just as well as the others. Difference in settings? The overall trend seems to suggest Anandtech's results are awry; the ones here and at [H] suggest a slight disadvantage at best at 16x12, not a slaughter.
Great review though -- compared several major brands!
[offtopic]
£1800? That's about 2,687.89! What do you plan buying? Core 2 Quad, GTX SLI? I paid roughly 1,300 for my PC (see sig)...
[/offtopic]