Mainstream Graphics Card Roundup

[JohnnyMash]

Page 15, "Current Nvidia graphics chips usually produce a CPU score between 6300 and 6400, whereas Nvidia chips come in around 6600."

Don't forget Nvidia chips.

 

[commisar]

LOL what an interesting article...
I have never been a fan boy for any company but i will say this, I own a single, reference, 4890 1 gig card (it was the best deal at the time) and i play crysis at 1920 by 1200 with everything on gamer settings and it runs like a dream. Farcry 2 @1920 by 1200 and everything turned up full and i never have a problem. the card is as solid as and i couldn't have asked for more... It was also alot cheaper than any of the nvidia cards around its performance bracket at the time of sale. ALOT!

 

[TheGreatGrapeApe]

[citation][nom]cangelini[/nom]Just an FYI guys, The Last Remnant was added to the benchmark suite after we received enough reader feedback asking for it--same with GTA 4. You'll likely see the suite vary from time to time based on the email and comments we're getting, but it certainly has nothing to do with the hardware being tested. And kudos to whoever it was that said TLR shouldn't be pulled just because ATI doesn't perform well in it. If anything, fans of the game should be clamoring for better driver support in it from ATI![/citation]

Still doesn't explain why DX10.1 was omitted in Hawx, and it doesn't garner the 'clamor to nVidia' comment, when the performance difference is there, as seen by other sites;

http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/16400-nvidia-geforce-gtx-275-896mb-review-19.html

Seems for balance you should tell Tino to do his next review WITH DX10.1 enabled, because people here have been asking for it long before they asked for Remnant to be added instead of Crysis.


Also, Enough with the delayed article dumping! It was bad before with old drivers, and getting worse! I'm sick of seeing EURO reviews that are posted here 3 weeks after being posted in Europe (May 29 in this case) when they already took 3+ weeks to test/write/vet/edit, etc. where even there they were getting criticism for being late.
http://www.tomshardware.com/de/Grafikkarten-Vergleich,testberichte-240338.html
IMO, either get it translated & up on the N.American site within 48 hours (plus your benefit of the time delay across the Atlantic) or else don't bother posting it , it reflects poorly on THG as being behind the times.

This is an article which could easily have been done by one of the North American review team, and based on the results, likely done better too, especially since it's about PRICES which should reflect the local market being shown the article.
Compared to other N.American news sources THG is becoming as timely as a magazine (all mine I canceled subscriptions to because they were 3-6 months out of date) with almost as long lead times as someone who has to set up a printer and then distribute hard copies via train/plane/automobile, what is going on with THG !?!
I see no reason THG can't get the articles out as fast as their competition, heck Russian sites get their reviews translated and to market faster than THG!

 

[alizarincrimson]

Why is left4Dead even used as a benchmark since the game caps it's framerate @60? seems kinda stupid to use as a benchmark.

 

[jcknouse]

any possibility you at Tom's can manage to get a 1G or 2G 4850x2 and benchmark it?

i'd like to see just how much difference there is between it and big brother 4870x2

thanks

 

[Guest]

Sadly, I started reading the article only to stop reading it 3 minutes later when it turned out that this article has a 512MB GeForce GTX+ posting consistently higher results than a similarly clocked with absolute same specifications 1GB GeForce GTS250... Way to go on delivering credible articles...

 

[werepossum]

I take no side on Tom's nVidia bias or lack thereof (and I would point out that AMD swept the recommendations for this month, as it should have), but I do agree that if ATI cards are equipped with DX10.1, the game supports it, and the ATI cards are faster in DX10.1, then the DX10.1 results should be used in place of the DX10 results for ATI because that's how it will be played. That's no different than refusing to test a 1GB HD4890 because nVidia's GTX275 (which, full disclosure, I happily own) only has 896MB. No one is going to willingly deny himself an advantage simply because the video cards he didn't buy don't have it.

 

[joey_sfb]

'The Last Remnant' is another example why I support ATI as I support open marketn and competition. As they are both good for us as a consumer. Nvidia has shown us that it will exerts its negetive influnce on developers that inevitablely compromise my choices and value as a consumer.

 

[joey_sfb]

[citation][nom]neiroatopelcc[/nom]ATI never could get drivers right. That's not just lately, that's been like this forever. They do eventually make them work, just not really performing well for the newest products. Nvidia isn't always better though. When their drivers work, it's all good, but especially with older hardware the drivers don't always work as advertised.[/citation]

Have you use any of ATI's RV770 serie? It works perfectly for me!

 

[FiresOfGreed]

Id like to note that I MAX TLR with a 4870x2 on my new rig. Quite honestly, the results in this article should be thrown out and redone. Otherwise Tom's Hardware will only prove itself to be a rather biased site.

 

[ariestrance]

All of these benchmarks are very helpful, but I want to know if anyone can give an accurate assessment of what the performance drop in FPS (average - across the board) would be on an AMD system with dual channel memory rather than triple channel, which is considerably slower. Keep the CPU as a high end AM3, keep the hard drive the same, have 4 gb vs. the 6 gb of RAM.

We all know the intel systems are fastest for the gaming - for an added price - but I'd really be interested to see what the performance loss is for saving those extra bucks. 10? 20 fps at the most in extreme situations?

Isn't the graphics card what makes the most difference? I imagine that with a high end AMD machine and that 4870x2, the frame frate drop couldn't be more than 10? Right. I don't know.

That's why I'm asking. Great article though.

 

[masterwhitman]

Wow, really? No HD4980, a card released months ago and prices at least that old? Come on guys; if this took you a long time to put together and these prices were current at that time, you really need to state it explicitly at the beginning, or just update the prices upon publication. For the record, I don't care about TLR benchmark; I use an HD4850 and can play the game @ 1080p just fine. Goes to show you that benchmarks are very much not real-world performance, something I love HardOCP's review process for.

 

[Guest]

Instead of whining like little girls about timeliness and perceived bias, can we focus on something that is indubitably a valid technical error (or a type-o)??? Why does the test system have 6 gigs of RAM and run a 32-bit OS? To fully utilize all of that memory (and the core i7, for that matter), a 64 bit OS should have never been in question!

I have a desktop that is nearly identical to the test system used here, which I find flattering, whose only difference is 12 gigs instead of 6, and I have a GF9400. I read this article researching a video card upgrade and I learned a lot. Timeliness is only partially relevant in that he reported on current pricing. If I decide to upgrade my desktop in 2 months instead of 2 weeks, the technical info will still be (mostly) valid.

The author is a PUNDIT, like Rush Limbaugh. He backs up his opinions with the facts as he has gathered them, like Rush Limbaugh. You are entitled to disagree with him, but that does not make him wrong! Like Rush Limbaugh or any other pundit, taking his word as gospel or automatically hating everything he says is stupid. He has presented the facts that he feels make the case for his opinion. Read the facts, check them if you like, and draw your own conclusion. Your opinion may not be the same as his. If you don't like a THG review, click over to cnet.com. If you don't like Rush, tune in to Air America. Slinging barbs from an ivory tower does nothing, even if you propel them with your own facts. Your facts will get lost in the venom of the name-calling, just as most pundits can legitimately dismiss most of their critics out-of-hand because they do little more than pitch insults backed by flimsy facts.

 

[Guest]

How do you use 6 GB ram with 32 bit operating system? And does it really make any sense to do so?

 

[Marcus52]

The tomshardware graphics card roundups are awesome, I link them to others with general video card questions (which is better? kind of thing) all the time. What I would LOVE to see though is a similar monitor comparison chart. Better even would be an article specifically on comparing TN to TN monitors - we all know it's the least color-accurate, but the fact is that's what most people are buying for their monitors and there's a huge price range amongst them. Which is better? What would you buy? Is TN really the hot ticket for twitch-gamers?

And why on earth would I spend a couple of grand for a monitor called professional when it's specs aren't any better than a $500 dollar one? What does 1 billion colors REALLY get you, as an enthusiast or professional, when that difference does exist. Is a $2k+ monitor actually more accurate than other IPS or PVA monitors)?

 

[Guest]

these tests are so biased and it just convinces me that Toms Hardware has lost all credibility and is not being objective at all...and i defy anyone to get s.m 2, s.m 3 scores as high as these obviously inflated amounts, as i have tested my 9800GT, 9800GTX+, and, my 9800GX2 cards on multiple systems and they come nowhere near these ridiculous figures....i deliberately said s.m 2 and 3. scores so my C.P.U's are not being factored in the equation and skewing my results as its just the graphics cards ability which i looked at which is irrepsective of my processors abilities, and still i have no idea where Tom is getting these ludicrous figures from....they are probably typos...and yes all my tests were done on clean installs of Vista with all updates and patches included, defragged etc, all running applications and anti virus switched off....9920 for s.m 3 for 9800GX2?? what a joke, who is they trying to kid?