Roundup: Mainstream Graphics Cards From ATI And Nvidia

[murillians]

the 9600 gso is also a good card with 96 cores

 

[awaken688]

I just wanted to add that I do enjoy having the power and NOISE. I care about performance, but I'll give up 5-10% performance if it will run quiet.

 

[scook9]

Well I know that I just got a 4890 for my HTPC (I game on it) for $150 after rebate, so I will be happy no matter what this review shows.

Got the HIS 4890 from microcenter but had them match the newegg price, $299 -> $199 then I still get the $50 MIR from microcenter

You all should jump on that while the rebate is good

 

[KT_WASP]

I am also scratching my head as to why the HD 4850 did not make the recommended list at the end, when they can be had for $100 with free shipping.

And no "We used prices at the time of the article" excuses. The HD 4850 have been at $100 for a good long time now.

I can't see how you can recommend the Zotac 9800 GT over the HD 4850 when your own charts show the HD 4850 wining the majority of the benchmarks (excluding Last Remnant), sometimes by 20+ frames. And that is at $35 cheaper...

Going by your methods to create that last chart $$$/fps , I decided to stack the HD 4850 against the Zotac 9800GT....

Zotac 9800GT : $135 | 961.1 FPS | $.140 per FPS
HD 4850 : $100 | 1040.5 FPS | $.096 per FPS

Zotac 9800GT : 296 watts under full load
HD 4850 : 270 watts under full load

Zotac 9800GT : 46.8 dB
HD 4850 : 47.9 dB

Zotac 9800GT : 87C under full load
HD 4850 : 94C under full load

I don't know about you... But I think the HD 4850 is clearly the winner...

 

[AdamB5000]

Cost per frames per second ($/fps) - I love it.

 

[hixbot]

4770 is in demand, and in short supply. If you can find one, you'll notice it costs more than a 4850. With that in mind, I don't know how you can, currently, recommend the 4770.

 

[Kithzaru]

Great article. Thank you toms.

 

[DarkMantle]

Third Tom's Roundup that uses The Last Remnant a game that ATI obviusly have problems with, again not using DirectX 10.1 on games that support it, no HD4870 1GB? come on, if you are going to use a GTX260(216)you should use the 1GB version of the HD4870 you used one on your other roundups and its the same price or cheaper than then GTX260. Recommending a 9800GT over a HD4850(clearly the winner of both) is a joke.

 

[dingumf]

[citation][nom]bloodblender[/nom]All I can say is that Tom's recent articles have been an excellent read, and this exactly the stuff I (as well as many others) require for their research purposes. Keep up the great work![/citation]

Research purposes? You are complete idiot.

Buying a GPU does not require much research. It all falls down to your budget. How much are you going to pay and what performance that price can achieve.

 

[scrumworks]

If you are curious to see how Radeons perform in HAWX using DX10.1 you can check it out e.g. here: http://www.guru3d.com/article/msi-radeon-4890-cyclone-soc-review-1ghz/12

 

[scrumworks]

Seems that Tom's fancy embed URL-link feature doesn't work (also can't even edit my comments) so here it is again in plain text which hopefully doesn't get filtered out this time: http://www.guru3d.com/article/msi-radeon-4890-cyclone-soc-review-1ghz/12

 

[KT_WASP]

I also want to point out that there is an error in the last chart.

You have the HIS H467QT512P as $95, but , then next line down, you then list two of them in crossfire for only $81. Clearly the Crossfire cost is wrong.

Is that card $95 each or $81 each?

It should read something like this...

HIS H467QT512P CF ($95 per card): $190 | 1169.1 FPS | $.163 $$$/FPS

or

HIS H467QT512P CF ($81 per card): $162 | 1169.1 FPS | $.139 $$$/FPS

With those new numbers, you can see how other, single card solutions, beat it out in the $$$/FPS category. Especially if it is $95 per card, as that would place it way to expensive for the frame rates your paying for.

 

[rambo117]

I thought that overall this was a great article. ive been waiting for a good one like this for awhile now.
probably gonna have to read it again, lots of information to take in all at once. keep it up tom's!!!

 

[Julianbreaker]

[citation][nom]amnotanoobie[/nom]The only reason that I could think of wherein the 4770 is better, is the smaller manufacturing process which should make it cooler and consume less power. Though if raw performance is your concern, the 4850 may be better.[/citation]

For what else would you need a graphics card for if not for raw performance? Would you actually be willing to spend an extra $10 for lower FPSs/folding points but a slightly more energy efficient rig? I can't imagine that being a viable solution even for an HTPC given the fact that who the dickens would stick a high-performance gaming card into an HTPC case anyway?

I mean, the case can be made that they weren't going to be comparing the 4850 other than as context but to actually go out on a limb and recommend a lower value card just because they just needed to recommend something given the number of GPUs they were reviewing is ridiculous.

That kind of logical fallacy seems to be why so many people (ie redditors) tell me to stop reading Tom's Hardware other than for their charts. Can't say that I'm likely to disagree with them.

 

[scook9]

[citation][nom]Julianbreaker[/nom]For what else would you need a graphics card for if not for raw performance? Would you actually be willing to spend an extra $10 for lower FPSs/folding points but a slightly more energy efficient rig? I can't imagine that being a viable solution even for an HTPC given the fact that who the dickens would stick a high-performance gaming card into an HTPC case anyway? I mean, the case can be made that they weren't going to be comparing the 4850 other than as context but to actually go out on a limb and recommend a lower value card just because they just needed to recommend something given the number of GPUs they were reviewing is ridiculous.That kind of logical fallacy seems to be why so many people (ie redditors) tell me to stop reading Tom's Hardware other than for their charts. Can't say that I'm likely to disagree with them.[/citation]


That would be me

Got a 4890 in my mATX HTPC (and it stays pretty dang cool while playing crysis too as well as quiet). As mentioned above, for $150 why not put in a 4890?! Got this card because it was the best that my case could fit haha

 

[Onus]

Provided that performance is acceptable, I will ALWAYS choose the lower-power option. I don't play anything really demanding like Crysis, but I'm currently testing a 4670 at 1680x1050 and I'm not suffering for not using my 4850. For now, I don't see a worthwhile difference. I'm glad to see some recent Tom's articles that have focused away from building the biggest possible edong onto what I would consider more rational builds.

 

[Shadow703793]

Interesting. If one already owns a 4670 it kind of makes sense to actually do CrossFire with these cards assuming one already has a CrossFire board.This would make more sense by the end of this year or beginning of next year as the way things are going you should be able to get a 4670 for little as $40-45.

 

[Onus]

Yes, I noticed that too; you could have a lot of graphics power without sucking a lot of juice or turning your PC into an oven.

 

[Summer Leigh Castle]

Can we get 4850 in crossfire? I would love to see the marginal difference between getting another 4850 versus moving up entirely to a new card.

 

[philosofool]

Nice focus topic. As a guy who will never, ever own a CF/SLI system or a $400 card, I'm glad to see an article focused on budget oriented cards.

However, Last Remnant and Tom Clancy are terrible bench mark games. Remnant is basically a console port that obviously doesn't play well with ATI. It would be one thing if this game were a rave review or huge seller like The Witcher, but why are you including a second rate game that doesn't work well with ATI cards as a bench mark? (It doesn't really matter that this game doesn't work with ATI cards, since your readers will just ignore the results. But it's an awful lot of work for you to do something that isn't worthwhile.) And Tom Clancy is capped at 30FPS, so its a pretty worthless game for high-end card bench marking. In short, you should choose games that give meaningful results and these two don't.

 

[Onus]

Any possibility that a pair of HD4670 cards in Crossfire would stress the power pathways of a typical mobo?

 

[pakardbell486dx2]

I just bought one of those HIS video cards with IceQ. IT IS NOT REAL COPPER!!!
I fond this out by accident. I was adding thermal grease to the core and i drooped the cooler and the "copper" block hit the ground and there was a slight dent on it. I noticed that there was two colors. To verify this i garbed a butter knife and scratch the heat heat sink. It was nothing but aluminum painted in a copper like color.
If Tom's has the time, they should grab on of those HIS IceQ and see for themselves.

 

[vizering]

"For what else would you need a graphics card for if not for raw performance? Would you actually be willing to spend an extra $10 for lower FPSs/folding points but a slightly more energy efficient rig? I can't imagine that being a viable solution even for an HTPC given the fact that who the dickens would stick a high-performance gaming card into an HTPC case anyway?"

Well, if the difference is $10.00 and you would opt for the cheaper, more power consuming and powerful card.

Would you not have to factor that into your monthly electricity bill and then multiply it by the amount of months you plan on using that card to see if it is cheaper, if at all?

 

[Julianbreaker]

"Would you not have to factor that into your monthly electricity bill and then multiply it by the amount of months you plan on using that card to see if it is cheaper, if at all?"

What a fantastic point completely unstressed by the conclusion of the article.

I will contend that, according to their PSU charts, that the power draw difference is negligible on idle and about 70 watts at full load. Since in most places you're charged something like 9 cents/1000watt hour, you'd be looking at a very long-term investment with the 4770 for value, which most likely will be an outdated card in a few years given its slower speeds in comparison to the 4850.

Again, it's not exactly an easy decision and I personally don't think that the 4770, right now at its current price point, deserves a recommendation.

 

[steiner666]

no crysis/warhead benchmarks = fail. who cares about how cards perform in games that don't stress them that much? I mean, seriously, L4D? FEAR2? FO3? ENDWAR?! Why dont you just add HL2 and DOOM3 to the games lists while you're at it. I think pretty much anyone who's bought a GPU in the past few years can expect to play those games with acceptable framerates. Very few ppl who are thinking about buying a new GPU will really be concerned with how it performs in lesser games such as these.