8800 ultra vs 250 GTS 1GB [Archive]

View Full Version : 8800 ultra vs 250 GTS 1GB

Vhlad

11-19-2009, 09:14 PM

Radiator

11-19-2009, 10:30 PM

Vhlad

11-19-2009, 10:50 PM

A GTS 250 is more or less equal to it at lower res , worse at higher res .
Now I wouldn't suggest a GTX260 to replace it , because the difference wouldn't be that MASSIVE . If you're going to replace it , replace it with a highend card , rather than a mid-end card .

I had to RMA a 8800 ultra. I kept it in pristine condition, always clean, cool, with no modification or overclocking, but it died on me after 2 years 4 months.

XFX mailed me a GTS 250 as a replacement after I asked them not to. :(

Their RMA policy states: if the same model card is not in stock a replacement of equal or better performance will be used as the replacement. They claim the GTS 250 is equal performance.

I'm curious how many people agree with that claim, and if so, what evidence they have to support it.

EVGA has been replacing RMA'd 8800 ultras with GTX 260s (i.e. http://www.nvnews.net/vbulletin/showthread.php?t=129748). So I'm a little bummed with XFX right now.

Radiator

11-20-2009, 03:15 PM

Vhlad

11-20-2009, 03:53 PM

Well thing is the GTS 250 IS more or less equal to it , it's the closest thing to a 8800 Ultra that's still on the retail market . The GTX260 is a notch upwards from the 8800U .
Don't fret , you can always sell the GTS 250 and buy a better card if it's not enough .

If the XFX RMA policy was written and advertised as: lifetime warranty, repair or replacement using the most equivalent card, sure. Because that would leave room for the replacement to be the closest thing, even if it were slightly lower performing.

But when they state: lifetime warranty, repair or replacement with equal or better performance, then they can't justify giving you the closest thing. It must be equal or better.

What I got was lower. :(
[at 1920x1200 with AA]

A slight performance difference is what separated the 8800 ultra and 8800 GTX at release. That slight performance gain came with a $200-$399 price premium.

Old prices are in this review: http://benchmarkreviews.com/index.php?option=com_content&task=view&id=34&Itemid=1&limit=1&limitstart=2

Old benchmarks are in this review: http://www.techarp.com/showarticle.aspx?artno=403&pgno=7

To quote the benchmark review: "the GeForce 8800 Ultra registered a 7.5% gain in frame rate over the GTX. This increased to 10% at 4x AA and 8xQ AA. The greatest amount of improvement was at 16xQ AA, where it was 12.6% faster than the GeForce 8800 GTX."

If it was a 8800 GTX being RMAd, then I would still be getting the GTS 250 from XFX. With the ultra being RMAd, I'm getting the same GTS 250 back, and I'm losing that slight performance gain that I payed so much for. i.e. GTS 250 1GB vs 8800 ultra, we're looking at 10.3% for sum of FPS benchmarks with 8AA; 17.8% for fallout 3]. It's like I never purchased the Ultra.

Radiator

11-20-2009, 06:55 PM

Well go there and demand a better card then .
At 1920x1200 or w/e the 8800 Ultra is indeed faster , but the GTX260 is way beyond the 8800 Ultra .
And I know what the 8800 Ultra cost when it was released .

-RK

11-20-2009, 10:37 PM

The GTS250 is slightly (very slightly) better.

Vhlad

11-23-2009, 03:01 PM

The GTS250 is slightly (very slightly) better.

I'd love to see benchmarks or other evidence to support this.

At stock speeds and resolution below 1920x1200, and no or little AA, I may agree. But on average, when ramping up the resolution and AA, the ultra outperforms.

I recently sent a support ticket in to EVGA to ask their opinion:
An EVGA support member wrote back, replying: The GTS 250 is a rebranding of the 9800 GTX super clocked. But the performance of the ultra cards are clocked a little faster than the GTX super clocked.

Aggravatingly, XFX didn't even send me the 1GB version of the 250 GTS to replace my 8800 ultra. According to the packing slip, it's a 250 GTS 512.

I'm not sure if I should bend over or just lay flat while they have their way with me.

-RK

11-25-2009, 01:04 AM

That is incorrect, the 8800Ultra has much better memory bandwidth due to a 384 bit bus, but the 9800GTX+/GTS250 has a significantly more efficient and powerful core.

It's technicalities, I know, but it remains that the 9800GTX+/GTS250 is to the 9800GTX what the 8800Ultra was to the 8800GTX. The ultra wins by a very slight margin at higher resolutions, the GTX+/GTS250 wins by a very slight margin at lower resolutions.

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Sum-of-FPS-Benchmarks-1680x1050,1197.html

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Fallout-3,1171.html

The operative words here are "very slight margin". The two cards are essentially equivalent in performance. No two card designs are going to perform in exactly the same way.

You can (and for your own sake probably should) raise a stink about it, but their logic is sound in sending you a GTS250.

Hey, the GTS250 should get you a better overclock too if you're into it.

ultima

11-25-2009, 01:33 AM

question

do you play at those resolutions ?

what size monitor ?

ultima

11-25-2009, 01:39 AM

also

did you LOOK at this comparison

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/compare,1199.html?prod[2464]=on&prod[2477]=on&prod[2465]=on&prod[2448]=on

yes some of the OLDER games it performs faster

however in NEWER games the 250 IS FASTER

look at fallout 3
far cry 2
HAWX

all beat the ultra

and to complain about a few frame rates
to me is ridiculous

when my video card of 3 yrs died
I HAD TO BUY A NEW ONE

getting lifetime warranty to me is silly for a company to do
its hows they go out of business

so are you going to be playing older games in the next year or two ?
or newer ones...

Vhlad

11-25-2009, 03:26 AM

I'm using a 2407wfp (dell ultrasharp 24"; native 1920x1200). Further replies are in darkred.

also

did you LOOK at this comparison

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/compare,1199.html?prod[2464]=on&prod[2477]=on&prod[2465]=on&prod[2448]=on

Yes, I have. Many of these comparisons are at low resolution (1680x1050) and/or without AA enabled, which means the inferior memory bandwidth of the GTS 250 1GB is not bottlenecking. If you look near the bottom of that link, you will see sum of FPS benchmarks 1920x1200, 8AA (high quality). Here you see the ultra beats the 250 GTS 1GB by 10.3%. This is not a single game or application, but the sum, yielding what can be interpreted as an average performance of each card. To say some things the GTS 250 wins and some things the 8800 ultra wins is technically correct, but to quantify that further you can look at the sum of FPS benchmarks.

yes some of the OLDER games it performs faster

however in NEWER games the 250 IS FASTER
On average, this is not correct (at 1920x1200 with AA).

look at fallout 3
I have. The 8800 ultra shows a 17.8% increase in performance over the GTS 250 1GB at 1920x1200 8AA 15AF
http://www.tomshardware.com/charts/gaming-graphics-cards-charts-2009-high-quality/Fallout-3,1511.html

far cry 2
1% performance gain for the GTS 250 (yes, just 1%); once again at 1920x1200 8AA 16AF (very high quality)
http://www.tomshardware.com/charts/gaming-graphics-cards-charts-2009-high-quality/Far-Cry-2,1515.html

HAWX
2% performance gain for the 8800 ultra
http://www.tomshardware.com/charts/gaming-graphics-cards-charts-2009-high-quality/Tom-Clancy-s-H.A.W.X,1530.html

all beat the ultra
Not true. The GTS 250 beats it by 1% in far cry 2, but the ultra performs 17.8% better in fallout 3, and 2% better in hawx.

and to complain about a few frame rates
to me is ridiculous
The difference between a 8800 GTX and a 8800 ultra was a few frame rates (12.6%); but the price difference between these cards for that gain was $200-$399. To care about a performance difference as high as 17.8% for fallout 3 or 10.3% over the sum of FPS benchmarks is not ridiculous to someone who bought a 8800 ultra. :)

when my video card of 3 yrs died
I HAD TO BUY A NEW ONE

getting lifetime warranty to me is silly for a company to do
its hows they go out of business

so are you going to be playing older games in the next year or two ?
or newer ones...
I'm playing fallout 3 right now. Why should I lose 17.8% in performance at my settings when RMAing a lifetime warranty card that I bought for $748.88 when the terms of the lifetime warranty specifically state that the company will repair or replace with an equal or better performing card (losing 17.8% is not equal).

It should also be noted that these benchmarks have been comparing the 1GB version of the 250 GTS. I was only sent the 512mb version.

I also like that link Ultima put up. But it uses a different version of the ultra. This one is more fitting:
http://www.tomshardware.com/charts/gaming-graphics-cards-charts-2009-high-quality/compare,1538.html?prod%5B2863%5D=on&prod%5B2856%5D=on&prod%5B2825%5D=on&prod%5B2807%5D=on
It compares 8800 ultra SLI vs GTS 250 1GB SLI and 8800 ultra single vs GTS 250 1GB single.

It's interesting: the SLI efficiency of the ultra appears to be better than that of the 250. i.e. the ultra SLI wins in 25/31 results (250 SLI performs a maximum of 3% better in its highest winning result, ultra SLI performs a maximum of 25% better in its highest winning result) and ties in 2/31 results; this is even though many of the results are at low res and without AA, where the 250 should do better.

Vhlad

11-25-2009, 03:47 AM

Thanks for the comments. I have a few quick replies in darkred.

That is incorrect, the 8800Ultra has much better memory bandwidth due to a 384 bit bus, but the 9800GTX+/GTS250 has a significantly more efficient and powerful core.

It's technicalities, I know, but it remains that the 9800GTX+/GTS250 is to the 9800GTX what the 8800Ultra was to the 8800GTX. The ultra wins by a very slight margin at higher resolutions, the GTX+/GTS250 wins by a very slight margin at lower resolutions.

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Sum-of-FPS-Benchmarks-1680x1050,1197.html
Well, there's a 0.9% difference in performance here, but it is at 1680x1050 and only 4AA. This resolution is not relevant to me, as my native resolution is 1920x1200.

http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Fallout-3,1171.html
The resolution is right this time (1920x1200), but it has 0AA and 0AF. We're talking about a performance difference related to the superior memory bandwidth of the ultra, and to see that performance difference the bandwidth must be utilized. i.e. ramp up the AA & AF. I was running a 8800 ultra SLI system. This was an enthusiast system without a doubt. People with these cards don't game without AA/AF. To be fair, AA/AF must be included.

The operative words here are "very slight margin". The two cards are essentially equivalent in performance. No two card designs are going to perform in exactly the same way.

You can (and for your own sake probably should) raise a stink about it, but their logic is sound in sending you a GTS250.
If they base their performance equivalency on low resolutions without AA, sure. But that's a questionable stance to take when we're talking about a card that was marketed toward people who cared about higher resolutions and AA.

Hey, the GTS250 should get you a better overclock too if you're into it.
I never overclocked my ultras. I thought with a SLI system I wouldn't risk it, due to the difficulty in finding a replacement that allows me to keep my SLI performance should anything happen. I was hoping something with a lifetime warranty kept at stock settings with stock cooler, air cleaned & well ventilated, would last long enough for keeping SLI performance to be a non-issue. i.e. I was hoping to have required an upgrade before one of the ultras died.

-RK

11-26-2009, 04:04 AM

The cards are equivalent. Raise a stink if you want, but it's pointless to try to argue that they aren't roughly equal.

You were using a SLI system and a $600+ card failed? Yeah, I can see why you'd be pissed to get a card that you can no longer use in SLI. Being pissed doesn't justify your argument or invalidate XFX's. The 8800Ultra is still slightly better at high resolution, slightly worse at low resolution. It sucks, yeah, but their position is still valid.

To be fair, they really should have sent you the 1gb version.

ultima

11-26-2009, 06:04 AM

yup should have been the 1gb version

unfortunately i wouldn't see them sending you a 260
cause it is a better card

maybe they'll send you a 4870

its faster than the 250 and a lot of the time the 260 too

and costs less than the 260..

maybe ??
;)

Radiator

11-26-2009, 03:51 PM

Highly doubt it... it's roughly equal to the GTX260 so I don't see why they should .

branewalker

12-17-2009, 07:08 AM

I have to say, the people here who are agreeing with XFX are not helping what is technically a very strong case for Vhlad.

Look at the benchmarks, then look at his system and the games he's playing. Oh, and the fact that the benchmarks are the 1GB card.

In his situation, especially, those cards are not equal.

I once worked for Best Buy. A woman came in one day looking to use her service plan to replace a 2+ year-old Sony Vaio laptop. Expensive, $2k (or more) machine when she bought it. Her service plan said, "equal or better specs" and the manager (thinking only of his immediate bottom line, and NOT of customer service) attempted to give this woman a Compaq Presario, the lowest-end laptop we had, as a replacement. He might have been correct if "specs" had meant "CPU speed and RAM." However, the original laptop was only 2.2 pounds, had a high-dpi screen (1280x800 at only 11 inches or so) and a few other goodies, like built-in camera, fingerprint ID login, etc. I (and many of my fellow customer service reps) made her case: in her situation, specifically, there was only one "equal or better specs" laptop, and it was the newer model of the same Sony Vaio. We ended up giving her an open-box unit (all we had in stock) of that model, but I think it was fair.

If they offer the warranty that says "equal or better" they better not stiff you by trying to rule out "better" with "almost equal." Equal would be another 8800 Ultra. Better appears to be the GTX 260. They guaranteed your card for their warranty period. You bought the first card because of that guarantee. If I were them, I'd be pretty certain that I had already lost a customer. I'd be bending over backward to make sure I got your business in the future, ESPECIALLY because you sound like the sort who would gladly pay $600+ for a NEW card someday, if he knew it was the best, and he could trust the company's warranty.

You're obviously and understandably disappointed.

I wouldn't "raise a stink," nor would I sue, but I'd certainly get on the phone with them and explain, exactly as you did here,
1. You are disappointed in their customer service
2. the card DOES NOT EQUAL your old one, as promised.
3. Tell them what would.
4. Tell them why. (Your 1920x1200 setup, benchmarks, 512 vs 1GB, etc)
5. Tell them how this will affect your future purchases.
6. Ask to speak with supervisor if they won't listen.

If it sounds like your're just trying to milk them on a technicality, they understandably won't care. If, however, this really affects 1) your gaming, and 2) your opinion of the company, then they ought to listen. And if they don't, well, you've learned what a warranty means at XFX.

Vhlad

12-24-2009, 01:42 AM

Thanks branewalker.

I wasn't originally planning to discuss my specific RMA issue on these forums. I was originally looking for general opinions on 8800 ultra vs GTS 250 performance.

I have a more detailed discussion of the RMA issue here:
http://www.hardwarecanucks.com/forum/video-cards/25699-disappointing-xfx-rma-experience.html

My final update in that thread (bottom of page 2) outlines the end results:
XFX still maintains: the GTS 250 is the replacement for the 8800 ultra, and that there is basically no difference between the 512 and 1GB versions of the GTS 250.

They did however take the 250 they sent me back, and yesterday I received a working 8800 ultra from them to conclude the RMA.

Radiator

12-24-2009, 01:42 PM

Well , cheers then .

-RK

12-25-2009, 01:14 AM

Ahh, it's good that you got another ultra. That lets you keep using your SLI setup, which should be better than even a GTX260.

I stand by my previous statements however. In general, the two are equivalent. The GTS250 is marginally faster at lower resolutions, the 8800Ultra marginally faster at high resolutions. In your case, the GTS250 is worse both because of the SLI setup and higher resolution you game at. Card manufacturers just can't realistically deal with things on a case by case basis though, which is why I recommended "raising a stink" (maybe that's an over-exaggerated euphemism).

However... it was wrong of them to imply that there isn't a difference between the 1gb and 512mb versions of the card, the fully enabled g92 core can easily make use of the extra RAM when it's needed.

ultima

12-25-2009, 01:45 AM

well its silly that they didn't send you that in the first
place

considering they magically found another one...

maleficarus™

12-26-2009, 02:09 AM

Hi GPUReview.

I have been comparing the 8800 ultra and the GTS 250:
http://www.gpureview.com/show_cards.php?card1=515&card2=606#

I was hoping to hear your opinion on the performance of these cards.

I hold the opinion that the 8800 ultra provides superior performance at 1920x1200 with AA/AF enabled. This is due to the higher memory bandwidth, and the higher pixel fill rate.

This opinion is also supported with benchmarks:

Sum of fps benchmarks at 1920x1200 with 4AA + high quality:
8800 ultra = 252.60
GTS 250 1GB = 247.70
http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Sum-of-FPS-Benchmarks-1920x1200,1199.html

When AA is increased further, so is the difference. Sum of fps benchmarks at 1920x1200 with 8AA + high quality:
8800 ultra = 121.70
GTS 250 1GB = 110.30
http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Sum-of-FPS-Benchmarks-1920x1200,1200.html

Looking at specific games instead of an average of all games, we see:
Fallout 3 at 1920x1200 8 AA 15AF very high quality:
8800 ultra = 44.90
GTS 250 1GB = 38.10 FPS
http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Fallout-3,1173.html

Left for dead at 1920x1200 8AA 16AF very high quality:
8800 ultra = 48.70
GTS 250 1GB = 43.80
http://www.tomshardware.com/charts/gaming-graphics-charts-2009/Left4Dead,1184.html

Strictly speaking, if a 8800 ultra replacement was required to be of equal or better performance, would the GTS 250 1GB suffice? Based on these benchmarks, and based on the superior memory bandwidth & pixel fill rates of the ultra, I believe that a GTX 260 would be required to meet the "equal or better" performance standard. What do you think?

Looking at your links there is something you have to understand right now. The human eye canot tell the difference between 252.60 fps and 247.70 fps nor can it tell the difference between 121.70 fps and 110.30 fps. This is not possible. The human eye can not see anything past 60 fps, this is why 60 frames per second has always been the industry standard ever since quake2 went online back in 1996. So the scores/links you are showing have no real backbone to support any real agument. Finally looking at the lower scores, they are so close that in the PC gaming industry they call that more or less equal. Anything 1% to 5% is equal.

This isn't like taking a car to the race track and one car can do the 1/4 mile in 10 seconds while the other car does it in 7 seconds and this is a big difference.

Having said that the GTS 250 is the better card for a few reasons. It is clocked using faster freqs, draws much less power from your PSU which by itsself should stand out as a better buy. For you to take older tech by judging it on frames that your eye will not ever actually see or notice was a bad, silly move my young guru jedi....

Vhlad

12-26-2009, 03:36 AM

Looking at your links there is something you have to understand right now. The human eye canot tell the difference between 252.60 fps and 247.70 fps nor can it tell the difference between 121.70 fps and 110.30 fps. This is not possible. The human eye can not see anything past 60 fps, this is why 60 frames per second has always been the industry standard ever since quake2 went online back in 1996. So the scores/links you are showing have no real backbone to support any real agument. Finally looking at the lower scores, they are so close that in the PC gaming industry they call that more or less equal. Anything 1% to 5% is equal.

This isn't like taking a car to the race track and one car can do the 1/4 mile in 10 seconds while the other car does it in 7 seconds and this is a big difference.

Having said that the GTS 250 is the better card for a few reasons. It is clocked using faster freqs, draws much less power from your PSU which by itsself should stand out as a better buy. For you to take older tech by judging it on frames that your eye will not ever actually see or notice was a bad, silly move my young guru jedi....

If you consider anything within 1% to 5% to be equal, then you would have no choice but to admit the GTS 250 1GB is not equal to the 8800 ultra at my settings. The two lower benchmarks from the post you quoted reveal: the 8800 ultra performs 17.8% better in Fallout 3, and 11.2% better in Left for Dead. Not only is that a substantially greater difference than 1% to 5%, but those results are also below 60 FPS (i.e. very meaningful). And, again, those benches are for the 1GB version of the GTS 250, whereas I was sent the 512mb version.

I try to explain my position further in subsequent posts. i.e. 8800 ultra SLI performs up to 25% better than GTS 250 1GB SLI. Thanks for taking the time to reply. If you're curious, in addition to the OP you should read the rest of my replies in this thread. There's also a post here describing the final resolution to the matter: http://www.hardwarecanucks.com/forum/video-cards/25699-disappointing-xfx-rma-experience-2.html#post306862

PS. The ultra has 47.3% higher memory bandwidth than the GTS 250. That's why it performs better at higher resolutions with AA.

feergoortens

03-15-2013, 12:52 AM

Propecia prevents the conversion of testosterone to dihydrotestosterone (DHT) in the body. Propecia is used for the treatment of male pattern hair loss on the vertex and the anterior mid-scalp area. Male pattern hair loss is a common condition in which men experience thinning of the hair on the scalp. Often, this results in a receding hairline and/or Buy Propecia Online Pharmacy (http://grou.ps/mdfedorova/428605) balding on the top of the head. Propecia is for use by men only and should not be used by women or children. Propecia may also be used for other purposes not listed in this medication guide. .

feergoortens

03-19-2013, 05:22 AM

What is Vigora?

A capsule was bewitched every 4 to 6 hours, 3 times per prime by way of healthful men and men with a luminosity erectile dysfunction. An obviousness amelioration of rigidity as grandly as with the ejaculation values bright and early are clearly shown to save each subject.
Conclusion Vigora does not earmarks of to show unfavourable effects on the blood pressure, cheapest canadian vigora online (http://measktwo.onlinewebshop.net/vigora/index.html), making its shoot up sheltered on all men. Posology and results. A capsule of 150 Mg was captivated during meals two (2) times per day, noontime and with supper. As shown in buy generic cialis 10mg (http://buycialis10mg.webs.com/buy-generic-cialis-10mg) the pigeon-hole beneath, before the intake of Vigora, 18 subjects (corps 1) presented a light erectile dysfonction; 6 subjects (guild 2) had a comprehensive erectile dysfonction with a view a crowning of 12 months epoch; 9 subjects (set 3) had a all-out erectile dysfonction in place of at least 8 years, including 2 subjects with implants; and 12 subjects (guild 4) were in perfected vigour without any erectile dysfonction.
Based on 4 days intake, 150 Mg Vigora has seemed ineffective on older men presenting a sum total erectile dysfonction quest of a want period. However, just out results show up that with doubling the amount, 300 Mg, older men with a total buy vigora no prescription (http://measktwo.onlinewebshop.net/vigora/buy-vigora-no-prescription.html) can look for a result with Vigora about bewitching 7 consecutive days intake (found up). For all man's categories, Vigora is a powerful nutraceutic countermeasure to boost waxing rigidity and delayedaction of ejaculation.

feergoortens

03-20-2013, 01:08 AM

Celebrex Weight Loss, Cheapest Celebrex, Celebrex Online (http://cmwp.onlinewebshop.net/celebrex/celebrex-online.html), Buying Celebrex

Celecoxib Price
Celebrex 20 Mg (http://cmwp.onlinewebshop.net/celebrex/celebrex-20-mg.html)
Buying Celebrex
Celebrex On Line
Celebrex Online
Celebrex For Sale (http://cmwp.onlinewebshop.net/celebrex/celebrex-for-sale.html)
Buy Celecoxib
Celebrex For Dogs 100mg