HD5770 (Sapphire Vapor-X 1GB) - so disappointing - Part 2

[Mussels]

5770's from sapphire have a problem with missing shaders, hindering their performance.


Oh and 3dmark 06 sucks really hard. cant beleive you returned a card based on that one benchmark.

 

[Kei]

Very true, it's pretty lucky to get your money back if something isn't broken. Whoever you bought it from is a pretty awesome company it seems.

Mussels, your post made me 'lol' really hard as I was typing mine.

Kei

 

[Peter_Palm]

5770's from sapphire have a problem with missing shaders, hindering their performance.


Oh and 3dmark 06 sucks really hard. cant beleive you returned a card based on that one benchmark.

That's ridiculous. I didn't returned the card because of that one benchmark.
I had serious other problems, too. I already explained that.

 

[Zubasa]

That's ridiculous. I didn't returned the card because of that one benchmark.
I had serious other problems, too. I already explained that.

Ridiculous as it sounds, Sapphire did mess up quite hard on that one.
There are already quite a few people that have the exact same card as you did and get it fixed.

 

[Kei]

Ridiculous as it sounds, Sapphire did mess up quite hard on that one.
There are already quite a few people that have the exact same card as you did and get it fixed.

Yes, I would've love to see the gpu-z readout on his card. I wonder if he was in the batch with the bad bios shader situation.

Kei

 

[Peter_Palm]

Not to heat the discussion again, but to clarify that also games exist, where the HD5770 is far away from winning the game (compared to a GTS 250):

"The GeForce GTS 250 performed admirably at 1920x1200, allowing us to use the highest available in-game graphics options. It was, in fact, the only video card in this segment to do that. Both the Radeon HD 5770 and HD 5750 required us to disable some in-game options to make Borderlands playable at 1920x1200.

With the Radeon HD 5770, we had to disable ambient occlusion and depth of field to make the game playable."

Source: http://www.hardocp.com/article/2009/11/30/borderlands_gameplay_performance_iq/5

And that happens with 1920x1200 pixels.

With 1680x1050, the GTS 250 is absolutely the winner.

See http://www.hardocp.com/article/2009/11/30/borderlands_gameplay_performance_iq/6

 

[crazyeyesreaper]

i prefer TPUs review and even with crappy drivers

http://www.techpowerup.com/reviews/HIS/HD_5770/8.html

a 5770 when the cpu bottleneck is removed it out paces the gtx 260 ati 4870 and matches the 4890 when overclocked

my guess is that your quadcore is holding the 5770 back as it seems most gpus these days like speeds in excess of 3.2ghz or so that seems to be the sweetspot

GTS 250 is still a damn good card for your resolution but the 5770 is still the superior card

and with good air cooling a q6700 should hit 3.5ghz which would really let the gpus stretch there legs so to speak

also you have to remember the gts250 is nothing more then a much higher clocked 8800gts basically 8800gts = 9800gtx = 9800gtx + = gts 250 thus drivers are heavily optimized by now to let the card do its work
so as drivers get better expect the 5770 to start walking away from the gts 250

but then again your running DX9 still so that makes a big difference as well

such as my 4870x2 when heavily overclocked could compete with my 5850s at stock but in dx10 the 5850s walk all over my older 4870x2 by massive margins same happens for you i bet that or when swapping cards drivers were not properly cleaned out which can cause performance issues eitherway tho the stock quadcore and DX9 are holding BOTH cards back at this point

 

[overclocking101]

it just seems that this guy prefers nvidia over ATI and thats absolutley fine. there are plenty of things that nvidia has the advantage over. I agree that your cpu is heavily bottlenecking your system and am kind of unsure why you dont have it running a bit faster. the Q6700's overclock quite nicely in fact and quite easily. the 5770 128 bit memory bus does hinder performance especially when being compaired to a card with a 256mb bus. apples to apples if you encode a lot of video etc and prefer the nvidia card keep it, but you really shouldnt come on and say its just straight better than another card and expect things not be heated and get flamed a bit. personally I prefer nvidia as ATI has left a bad taste in my mouth.

 

[crazyeyesreaper]

well everyone has preferences and at 1440x900 the GTS 250 is still extremely capable i just prefer newer tech myself thus the GTS250 left a bad taste in my mouth ill admit ATis driver do tend to suck especially at launch but same can be send with nvidia and certain games flat out not running on certain cards due to driver issues.

end of the day tho either a 5770 or a gts 250 will be much faster if he clocks up that quadcore let the card stretch its legs and go to town lol

 

[Zubasa]

Not to heat the discussion again, but to clarify that also games exist, where the HD5770 is far away from winning the game (compared to a GTS 250):

"The GeForce GTS 250 performed admirably at 1920x1200, allowing us to use the highest available in-game graphics options. It was, in fact, the only video card in this segment to do that. Both the Radeon HD 5770 and HD 5750 required us to disable some in-game options to make Borderlands playable at 1920x1200.

With the Radeon HD 5770, we had to disable ambient occlusion and depth of field to make the game playable."

Source: http://www.hardocp.com/article/2009/11/30/borderlands_gameplay_performance_iq/5

And that happens with 1920x1200 pixels.

With 1680x1050, the GTS 250 is absolutely the winner.

See http://www.hardocp.com/article/2009/11/30/borderlands_gameplay_performance_iq/6

So the GTS 250 is faster in Game X so it must be the better card despite the fact that the HD 5770 is much faster on average?
Nice reviews with every card running @ different settings.
Those reviews aren't even meant for comparing different cards.

 

[DirectorC]

Those reviews aren't even meant for comparing different cards.

They might not be MEANT for that, but reading several reviews of a wide range of products pitted against each other is actually a great way to compare those products.

 

[Zubasa]

They might not be MEANT for that, but reading several reviews of a wide range of products pitted against each other is actually a great way to compare those products.

The thing is these review employ 16xQ CSAA...:shadedshu
So we should turn on CFAA so that we can compare apples with oranges?

 

[Peter_Palm]

So the GTS 250 is faster in Game X so it must be the better card despite the fact that the HD 5770 is much faster on average?
Nice reviews with every card running @ different settings.
Those reviews aren't even meant for comparing different cards.

Yes. That's why the content behind the link with the 1920x1080 test results has to be corrected - but if that would be done, the HD5770 (and the HD5750) would get even more worse results... I guess that's not what you want/would expect.

 

[bobzilla2009]

Well, okay, if you say so. But, DX11 seems very promising. Just look at Dirt 2 or the Heaven benchmark and the difference between them is quite noticeable. Plus, Windows XP is over 8 years old anyway.

I tried Windows XP (x64) on my Phenom II X4 and performance was pretty choppy. Windows 7 has better support for multicore processors, as Windows XP doesn't seem to work well with more than 2 cores. Windows 7 x64 simply flies on my machine. Plus, I enjoy the new GUI over the old 2D, unbuffered GUIs of yester-decade. Not sure why people complain so much about the GUI (unrelated to this topic, though.)

Windows xp is horrible beyond belief for anything more powerful than a toaster in the processor department. However it can do games ok, mostly anyway. Really, if anyone is a pc gamer/ enthusiast, you owe it to yourself to move to windows 7 if you have xp

 

[crazyeyesreaper]

all i know is in almost every other review out there the 5770 is the better card both tech wise and performance wise even in games heavily favorable to Nvidia it holds its own

not to mention the review you posted is the only review ive seen where the gts250 is on top guru 3d techpower up legit reviews all show the same thing the 5770 is the better card

but on win xp with that lower clocked quad you wont see any real improvment windows XP had its glory day when the 8800 series was king as it stands now its on its last legs driver support is shady and its just plain time to move on

 

[Mussels]

not every one of these cards has the issue, it may well be a small batch as they stated. I contacted a few of my favourite stores here in Au, and i keep getting emails back "thanks for letting us know, we contacted sapphire and our batches are unaffected"

 

[Peter_Palm]

all i know is in almost every other review out there the 5770 is the better card both tech wise and performance wise even in games heavily favorable to Nvidia it holds its own

not to mention the review you posted is the only review ive seen where the gts250 is on top guru 3d techpower up legit reviews all show the same thing the 5770 is the better card

but on win xp with that lower clocked quad you wont see any real improvment windows XP had its glory day when the 8800 series was king as it stands now its on its last legs driver support is shady and its just plain time to move on

I am still convinced it depends on the indended use.
I have tested my (overclocked) GTS 250 with another, different game benchmark and it still looks very good, see http://www.z80.eu/images/farcry2_bench_1280x1024.jpg .
So for me the 5770 is not always the better card.

 

[InTeL-iNsIdE]

Hi,

just to complete my tests:
I've changed my card again to a GTS250 (MSI N250GTS Twin Frozr 1G), which is oc'ed by factory (GPU driven with 760MHz instead of 738MHz).
My HD5770 was also overclocked by factory to 860 MHZ (which is a very small amount).

Both was oc'ed manually to a higher value:
The HD5770 to 900MHz GPU frequency,
The GTS250 to 770MHz GPU frequency (this is not the highest amount, I was able to get it up to 780MHz without errors, but I want to have a stable system)

Here is the 3DMark 2006 result of both:
http://www.z80.eu/images/GTS250vsHD5770.jpg

It's only a difference of ~500 points, but also a difference of $50 (the GTS250 is cheaper).
Judge for yourself.

P.S.: And no more problems with DualView (2 monitor) flicker, and no problems anymore with Star Wars Battlefront II ...

your being held back by your q6700, 3d06 is notoriousley cpu based and not representative of games. A 5770 sits between a 4870 and a 4890 in terms of performance, not too mention has dx11 and also uses less power. You have successfully downgraded gpu's, give yourself a pat on the back!!

 

[Peter_Palm]

your being held back by your q6700, 3d06 is notoriousley cpu based and not representative of games. A 5770 sits between a 4870 and a 4890 in terms of performance, not too mention has dx11 and also uses less power. You have successfully downgraded gpu's, give yourself a pat on the back!!

What elitist thinking. For me, I managed it with a new CPU cooler to get 3 Ghz for each of the 4 cores. That's above average already. What do you think what a typical gamer should invest in his hardware ? An i7 is far away from my expectations from a good value for money.
And I have also to "sponsor" other things as well, I do not live and work just for my PC.
Also, all my games are now running with a lot of reserve already (in terms of fps). So for me it's not really important to get 120fps instead of "only" 90fps (it's still more than enough).
I just wanted to point out that even after 2 years of no really new developed hardware from nVidia, it's still competitive. And I like to have drivers without glitches/errors. ATI does not reach such a stability yet.
And DX11... lol... I already mentioned here that without games for DX11, there is no need yet for this.

 

[InTeL-iNsIdE]

What elitist thinking. For me, I managed it with a new CPU cooler to get 3 Ghz for each of the 4 cores. That's above average already. What do you think what a typical gamer should invest in his hardware ? An i7 is far away from my expectations from a good value for money.
And I have also to "sponsor" other things as well, I do not live and work just for my PC.
Also, all my games are now running with a lot of reserve already (in terms of fps). So for me it's not really important to get 120fps instead of "only" 90fps (it's still more than enough).
I just wanted to point out that even after 2 years of no really new developed hardware from nVidia, it's still competitive. And I like to have drivers without glitches/errors. ATI does not reach such a stability yet.

So sick of hearing the same shit from nvidia fanboys, I have been using ati for the last 2-3 years (before that i shelled out £180 for a g80 8800gts 320mb fyi) I have never had any issues with their drivers, I am not saying it doesnt happen, sure it does, but it also happens with nvidias drivers too, only its always nvidiots who always bring up ati having shitty drivers.

FACT a gts 250 is bested by a 5770 in almost every way and game unless you like suporting nvidia and only play twimtbp games, which from your trolling posts really wouldnt surprise me.

And yes a gts isnt THAT far off a 5770 (overall in all resolutions the 5770 is at least 15-20% faster btw) not bad for 2 year old rebranded tech, but as has been mentioned before it is basically a rebadged 8800gt/9800gtx which were both high end offerings, so really kudos to ati for their mid range beating nvidias last 2 gens high end.

Like I said, you have down graded, nvidia would be really proud of your support if they gave a dam.

 

[Peter_Palm]

So sick of hearing the same shit from nvidia fanboys... [...]

Now it's clear who is the fanboy.

 

[ToTTenTranz]

I'd say this thread and its OP have the only purpose to create a flame war, and should be stopped right away..


First the OP says he switched a HD5770 for a GTS250 because of a 4 year-old benchmark, then he states that it wasn't just because of that, there was also the hardocp benchmarks with different settings and resolutions between the cards..

Everything is just stupid..

Next time, I'll come here to say I switched my HD5870 for a GTS 210, because the former didn't support CUDA. Because I want to use CUDA to encode my videos, although I have a core i7 overclocked to 4GHz.

 

[Peter_Palm]

I'd say this thread and its OP have the only purpose to create a flame war, and should be stopped right away..


First the OP says he switched a HD5770 for a GTS250 because of a 4 year-old benchmark, then he states that it wasn't just because of that, there was also the hardocp benchmarks with different settings and resolutions between the cards..

Everything is just stupid..

Next time, I'll come here to say I switched my HD5870 for a GTS 210, because the former didn't support CUDA. Because I want to use CUDA to encode my videos, although I have a core i7 overclocked to 4GHz.

Ok, promised, I will stop saying my free opinion. Also, because of those ridiculous, unpolite and wrong quoted entries.
I even published numbers from FarCry 2 Benchmark, and if you read the hardocp article exactly, there was no different resolution taken, just a few options were switched off for the 5770 because otherwise the game was not playable anymore - so they had to switch the options on to show up how slow the 5770 is in that game (compared to the GTS250).
ATI Fanboys like you, feel free to go on dreaming about a good graphics card. The 5770 isn't such one, may be the 5850 and the 5870 are those.

 

[Zubasa]

Yes. That's why the content behind the link with the 1920x1080 test results has to be corrected - but if that would be done, the HD5770 (and the HD5750) would get even more worse results... I guess that's not what you want/would expect.

Its is completely pointless to say this.
There is no IF, I only believe in actual test results.
After seeing enough test results, you will find out that things are very often not as expected.

Ok, promised, I will stop saying my free opinion. Also, because of those ridiculous, unpolite and wrong quoted entries.
I even published numbers from FarCry 2 Benchmark, and if you read the hardocp article exactly, there was no different resolution taken, just a few options were switched off for the 5770 because otherwise the game was not playable anymore - so they had to switch the options on to show up how slow the 5770 is in that game (compared to the GTS250).
ATI Fanboys like you, feel free to go on dreaming about a good graphics card. The 5770 isn't such one, may be the 5850 and the 5870 are those.

Turning off certain effects don't always gives the 5770 an actual adventage over the other cards.
You really can't prove any point as long as the site you quoted runs their tests at different settings.
Those are simply not done under a standard enviroment.

 

[ToTTenTranz]

ATI Fanboys like you, feel free to go on dreaming about a good graphics card.

Really, how can anyone take this guy seriously?

He's here to start a flame war, obviously. Am I the only one seeing that?