DailyTech - 4/2/2009 Daily Hardware Reviews -- NVIDIA GeForce GTX 275 & ATI Radeon HD 4890 Edition

Submit News

Hardware 4/2/2009 Daily Hardware Reviews -- NVIDIA GeForce GTX 275 & ATI Radeon HD 4890 Edition
Brandon Hill (Blog) - April 2, 2009 6:45 AM

9 comment(s) - last by legoman666.. on Apr 3 at 8:31 AM

ATI Radeon HD 4890 vs. NVIDIA GeForce GTX 275 @ AnandTech

DailyTech's roundup of hardware reviews from around the web for Thursday

NVIDIA GeForce GTX 275
@ AnandTech
@ PC Perspective
@ Guru3D
@ Driver Heaven
@ Hot Hardware
@ Hardware Canucks
@ NinjaLane
@ Bjorn3D
@ Legit Reviews
@ Overclockers Club

ATI Radeon HD 4890 Edition
@ TweakTown
@ Club Overclocker
@ TweakTown
@ Hardware Canucks
@ Hot Hardware
@ Benchmark Reviews
@ Overclockers Club
@ Elite Bastards
@ Hartware.de
@ AMD Zone
@ Bjorn3D
@ InsideHW

Processors
AMD Phenom II X3 720 and X4 810 Processors @ iXBT Labs

Motherboards
Biostar TPower X58A v5.0 @ Trusted Reviews
ASUS Rampage II GENE @ Bjorn3D

Storage
In Win ‘NA3.5″ Hard Drive Enclosure @ CCE Reviews

Peripherals
Razer Arctosa Gaming Keyboard @ Tweaknews

Cooling
Coolermaster N520 Dual Fan CPU Cooler @ Pro-Clockers
Thermaltake SpinQ CPU Cooler @ Pro-Clockers

Power Supplies
PC Power & Cooling Silencer 610W PSU @ CPU3D
NZXT Performance Power 800W Power Supply @ ThinkComputers.org
Nexus Value 430 Watt (Eco) PSU @ XtremeComputing

Consumer Electronics
Nikon CoolPix L19 and L20 @ Trusted Reviews

Comments

Threshold

Username
Password
remember me

This article is over a month old, voting and posting comments is disabled

Anand review disappointing

By Warren21 on 4/2/2009 9:51:05 AM , Rating: 5

I like the coverage they did about PhysX and CUDA, it was nice to see them include those as part of the consideration for the GTX 275.

However... They play the HD 4890 off as being an OC'ed RV770 and don't even mention the fact it uses a slightly larger, different RV790 die with tweaked power phases, efficiency... etc. Other reviews have showed the new card as having lower idle consumption than the 4870 1GB while retaining the higher clocks. Not to mention the fact it clocks to 1 GHz core stable on stock cooling? I can't think of a card that has ever done that, and I'd consider it a major selling point (keep in mind this is through CCC overdrive... It's not even hard to do).

Finally, later on in the article they compare the cards amongst others in bar charts where you can clearly see the 4890 leading the GTX 275 at all but 30" panel resolutions, yet they claim that "It looks like NVIDIA might be the marginal leader at this new price point of $250." Who do you know that buys a 250 USD card to play at 2560 x 1600? Seriously guys? Winning at 2/3 tested resolutions (1680 x 1050 and 1920 x 1200) seems like more of an advantage for me.

All in all, decent but disappointing.

RE: Anand review disappointing

By lakrids on 4/2/2009 10:11:53 AM , Rating: 3

quote:
Not to mention the fact it clocks to 1 GHz core stable on stock cooling? I can't think of a card that has ever done that, and I'd consider it a major selling point

Agreed.
insidehw, bjorn3d, amdzone, hardwarecanucks and xbitlabs ALL have their cards hitting 1ghz.
Other sites like hardocp, hothardware, overclockersclub, hartware.de hit decent ranges of 970-985 mhz.

Amazing overclocking capatibilies adds value. And I think the RV790 does this decently, RV770 is nowhere near this 1ghz so ATI/AMD must have done something right.

Parent

RE: Anand review disappointing

By therealnickdanger on 4/2/2009 10:35:38 AM , Rating: 2

Yeah, I was disappointed that they didn't touch on its OC capabilities. I was also confused that they gave the 275 the kudos they did considering the 4890 seemed to whoop its butt in nearly everything.

I will say that when ALL things are considered and if pricing/performance are equal, the 275 is the better value due to its extra features - even if those features have nothing to do with fps. There is an overall hostile tone toward NVIDIA lately - not that they don't entirely deserve it - but it seems to cloud some basic realities.

Parent

RE: Anand review disappointing

By LuxZg on 4/2/2009 11:16:46 AM , Rating: 2

I agree with this post by Warren21, I was just browsing through article as I've already read 4-5 articles about HD4890, and was watching 1920x1200 in particular. ATI card won the majority of benches, and yet in conclusion ther was this "GTX275 is slightly better" talk. I second the original poster, those using 30" screens will have enough money to buy way better (dual-GPU) cards, or use HD4890 in SLI. 1680 and 1920 are main target for buyers of <250$ cards, so it should carry more weight than higher resolution(s).

As for this DailyTech article bringing bunch of links, there is a huge miss in the whole story. Big chunk of those articles talking about one card, have results of the other card as well, and treat them pretty much equaly in reviews. Yet some are sorted in HD4890, and some in GTX275 "section", while section named "both" is severely lacking. Not to mention that most of the articles listed in GTX275 section actualy have HD4890 as main card on review (both in title of article, and written about more extensively). So please, at least fix this small link-list, even if you don't change the Anandtech reviews conclusions.

Parent

RE: Anand review disappointing

By silversound on 4/2/2009 3:31:14 PM , Rating: 3

(1680 x 1050 and 1920 x 1200) should be medium to high resolution, 2560 x 1600 is ultra-high. I doubt very few people use a 30" monitor to game, mine 24" already too big for me :p

Parent

flip the pictures

By kevinkreiser on 4/2/2009 9:43:30 AM , Rating: 2

you may want to flip the pictures. the ati text is next to the nvidia card and vice versa. just a suggestion :o)

RE: flip the pictures

By LuxZg on 4/2/2009 11:54:52 AM , Rating: 1

Actually, they should swap a lot more than just pictures :/

Let's see what these links (in order) are writing about:
- Anandtech has title "ATI Radeon HD 4890 vs. NVIDIA GeForce GTX 275" so it's actualy BOTH (even though 8 pages are about PhysX/CUDA, like anyone cares THAT MUCH for those 9 games in total and 2 applications that use CUDA)
- PCPerspective's title is "AMD Radeon HD 4890 RV790 and GeForce GTX 275" - so again BOTH cards in it
- Guru3D has title "Radeon HD 4890 review | test" - so it's actualy HD4890 review, GTX275 isn't even mentioned in article
- driverheaven.net title - "ATI Radeon 4890 and NVIDIA GTX 275 Graphics Review" - so BOTH again
- Hot Hardware's link is finaly in right section as title of first link is really "NVIDIA GeForce GTX 275 Unleashed" - wow, you got one right :) Even though I'd put it in non-existant "both" category because they have HD4890 review as well, and they have numbers of HD4890 card in this review, and other way around
- Hardware Canucks is also in right category with title of article being "Nvidia GeForce GTX 275 896MB Review", but similary, it has scores of both cards in it, as they too have two separate articles (one for each card) - so I'd put it in "both" category as well
- NinjaLane article is first GTX275-only article, as it's both named "GeForce GTX 275 Video Card Review" and has no HD4890 in it at all, so finaly something for GTX275-only section..
- Bjorn3D again has 2 articles, as did Hot Hardware and Hardware Canuks, and both have numbers for the "other" card
- Legit Reviews is naming their article "NVIDIA GeForce GTX 275 Versus ATI Radeon HD 4890" so why it's in GTX275 section? Again, should be in imaginary "BOTH" section of links
- Overclockers Club again has both articles, with all cards in it, oh, and they have 2 HD4890 articles, one isn't even listed on this page ( http://www.overclockersclub.com/reviews/sapphire_h... )

Now we're off to HD4890 section of links, and things are much clearer here, finaly. Oh, except articles already mentioned above, ofcourse. Others don't include GTX275 in their reviews so it's fine here. Except the picture showing nVidia card like Kevin already mentioned ;)

So again, why not reshuffle, reorder and describe links a bit better. OK, if they primarily talk about one card, put them in proper section for that card, but at least put an asterisk (*) by it so we know that the other card is in graphs as well (and state under what "*" means :) ). And really, you should move some of the articles to the right section (like Guru3D article), and you should open a "BOTH" section for articles that did review of ATI vs nVidia cards all in one article, as Anandtech's for example.

Also, you lack several good links there..
Xbit labs - http://www.xbitlabs.com/articles/video/print/radeo... (this is HD4890-only article)
Tomshardware - http://www.tomshardware.com/reviews/radeon-hd-4890... (this one is about HD4890 ONLY!)
http://www.tomshardware.com/reviews/geforce-gtx-27... (this one is about GTX275, but numbers from HD4890 article are also in it)
Firinsquad - http://www.firingsquad.com/hardware/ati_radeon_489... (talks about both cards, even has 4890 in crossfire)
Gamestar.de - http://www.gamestar.de/hardware/tests/grafikkarten... (just about HD4890)

Oh, btw, that hartware.de article has huge list of cards in their article's graphs! :-O Great to see what you're getting if you're upgrading from say.. X1950XTX :) Even though selection of games is rather.. old :)

Parent

RE: flip the pictures

By just4U on 4/2/2009 5:43:11 PM , Rating: 2

What's interesting about the firingsquad article is temperatures. They took AMD to task on 4870 ones and yet.. the Sapphire card I have has indentical temps to the new 4890. Im guessing Sapphire had already corrected these issues as some people seem to complain about temps and some say their totally fine (on the 4870 I mean)

The Nvidia temps are really nice to.. glad to see that included because I always want to know if Im getting a heater or not (like in the 3850 and 9800GTs)

Parent

RE: flip the pictures

By legoman666 on 4/3/2009 8:31:06 AM , Rating: 2

High temps do not necessarily mean it has a high heat output. You could heat a penny up to 200F with 5w of power if you didn't cool it. You could also keep the penny at 110F even while applying 500w of heat to it if it was cooled properly. If you're simply interested in electricity usage and heat output, then look at those numbers; the temp has nothing to do with it.

Parent