qubit
Overclocked quantum bit
- Joined
- Dec 6, 2007
- Messages
- 17,865 (2.93/day)
- Location
- Quantum Well UK
System Name | Quantumville™ |
---|---|
Processor | Intel Core i7-2700K @ 4GHz |
Motherboard | Asus P8Z68-V PRO/GEN3 |
Cooling | Noctua NH-D14 |
Memory | 16GB (2 x 8GB Corsair Vengeance Black DDR3 PC3-12800 C9 1600MHz) |
Video Card(s) | MSI RTX 2080 SUPER Gaming X Trio |
Storage | Samsung 850 Pro 256GB | WD Black 4TB | WD Blue 6TB |
Display(s) | ASUS ROG Strix XG27UQR (4K, 144Hz, G-SYNC compatible) | Asus MG28UQ (4K, 60Hz, FreeSync compatible) |
Case | Cooler Master HAF 922 |
Audio Device(s) | Creative Sound Blaster X-Fi Fatal1ty PCIe |
Power Supply | Corsair AX1600i |
Mouse | Microsoft Intellimouse Pro - Black Shadow |
Keyboard | Yes |
Software | Windows 10 Pro 64-bit |
Some of you asked for a performance comparison between the current Pascal based bottom end GT 1030 graphics card and the old Fermi based top end GTX 580 from 2011, so here it is. @P4-630 @Artas1984 @Frick @Prince Valiant this is especially for you!
I've benched them using Unigine Heaven 4.0 and the tl;dr version is that the GT 1030 has about 55% of the performance of the GTX 580 in DX9 and DX11 modes. Not too bad for such a low end card, even with the large generational gap between them.
Now, with that out of the way, I thought I'd have a bit of a benchathon this weekend, so expanded the number of cards to test, all of which I own. I have several more cards than this, but there has to be a limit somewhere, lol.
Here are the contenders, with TPU review links where possible. The review may not be of the same exact model, but will be functionally identical.
Zotac GTX 1080 Amp! Extreme Edition 8GB
MSI GT 1030 OC Edition 2GB (with passive cooling)
MSI GTX 780 Ti Gaming 3GB
Zotac GTX 580 Amp! Edition 1.5GB (reference cooler)
Palit GT 520 2GB (with active cooling)
Zotac GTX 285 1GB (reference cooler)
XFX 8800 GTX 768MB (reference cooler)
Just a note about the older cards not working. I had trouble getting the much older 8800 GTX and GTX 285 to actually give a picture, although the PC gave the POST beep and Windows booted otherwise. Turns out the gold PCIe connector pads had developed some visible tarnish on them which prevented good contact being made. Cleaning them with isopropyl alcohol fixed the problem, with the cards working perfectly after that. One to watch out for if your card plays up.
DRIVER
The oldest cards, the 8800 GTX and GTX 285 are no longer fully supported by NVIDIA so were run with the latest available Windows 10 64-bit driver, version 342.01. All the rest were run with the very latest version, 384.76.
The test rig is the one in my current system specs:
I used Unigine Heaven 4.0 in DX11 and DX9 modes. Since the 8800 GTX and GTX 285 only support DX10, I had to drop down the DX version for it to work. I chose DX9, because that was widely used, while hardly anything was released using DX10 and hence not representative of real world performance.
I used the Ultra quality settings, as shown:
DX11 with Normal tessellation level
DX9
RESULTS
Note that it wasn't possible to run the tests with ambient occlusion switched off for some reason. I would have preferred not to use it, because in my opinion it doesn't add that much to picture quality while hammering performance, so it's not worth it. This was especially true with the weaker cards and made them look a lot worse than they were otherwise. Dropping the quality down to Medium helped a lot, too.
One thing to strike me immediately, is that all the DX11 cards gave better performance in DX11 mode, even with tessellation on. With tessellation off the performance was considerably better, putting it significantly ahead of DX9. This is in line with the efficiency improvements touted for DX11.
Unsurprisingly, the GTX 1080 totally dominated all the other cards, with the GTX 780 Ti coming in a distant second place, having about 62% of the performance. This is still enough for good performance in many games where the 3GB RAM limitation isn't a problem.
The GTX 580 made a reasonable job of smoothly rendering the scenes at these punishing settings, with many of them being rendered a lot faster and hence smoothly than the average 45-50fps seen here would suggest. It did it with a lot of huffing and puffing though with that noisy reference cooler! Killing ambient occlusion would have improved the performance considerably, resulting in perfectly smooth animation throughout.
The GT 1030 gave about 55% of this performance with the animation looking decidedly stuttery, but just about playable at a push. Again, just killing that pesky ambient occlusion would have resulted in markedly better performance, offering reasonably smooth framerates at about double the speed. Putting the quality setting to Medium would then result in very respectable performance indeed.
So, if you have a GTX 580 don't replace it with a GT 1030 expecting better framerates. Only do so maybe if you're not playing demanding games and want a quieter card that consumes much less power and has newer features. Even then, at £65 to £70 it's not exactly cheap right now, so you're probably better off going with a GT 1050 if you can, which offers much better performance.
It's interesting to see how the legendary 8800 GTX from waay back in 2006 was unplayable at these settings although it was a DX9 powerhouse back then. It even supported the brand new DX10, for which there was next to nothing that used it at the time. Shows just how far the mighty have fallen! Even without ambient occlusion, it wasn't really up to much. Hit it with some older games though and it will still rock.
The GTX 285 was about 50% better, but at these framerates still useless. Back in its day it felt like a supercharged 8800 GTX.
And now we arrive at the runt of the litter: the little GT 520. At roughly 2.6fps, it was a full-on slideshow and so much so that the application's interface was hammered too, with me barely able to select the options. Turning off ambient occlusion gave all of 4-5fps. Wow. This means that its big brother of the same generation, the GTX 580 is about 20 times faster, a staggering difference! And the GTX 1080 completely embarrasses it with 70 times the performance.
Comparing the two bottom end cards, the GT 1030 and the GT 520 showed just how far low end cards have come, with the GT 1030 having about 11 times the performance. For anyone thinking that these low end cards have stagnated, think again - Pascal is potent even in its smallest form. Nowadays they have to beat half decent integrated graphics processors, so they can't be too low performance.
NOISE PERFORMANCE
Of course, another important parameter is the noise that these cards make when running a stress test like this and boy were some of them loud!
The 8800 GTX and GTX 580 where by far the noisiest, being really obvious about the effort they were making, with the 8800 GTX sounding like it was going to take off. The GTX 285 was still pretty loud, but better. All of those cards were still reasonably tolerable however, because there was just a lot of windrush noise rather than any irritating buzzes or whines. Credit to NVIDIA for thinking about this aspect of performance. Note that they all sound a lot quieter when normal gaming.
Surprisingly, the GT 520 was pretty quiet, possibly because it may not have a temperature controlled fan and the GPU was cooking...
The quietest card was unsurprisingly the passive GT 1030, which didn't even exhibit any perceptable coil whine and worked with total stability under heavy stress. Impressive. I wouldn't risk Furmark on it though with that passive cooler. It was pretty hot to the touch afterwards though, as were the rest of the cards.
The GTX 780 Ti made a clearly audible noise, but still pretty good for a card being stressed to the max, showing what a good non-reference cooler can do. I can tell you that in normal gaming this card is very quiet indeed.
Finally, the GTX 1080 was whisper quiet, taking all this in its stride, not making any more noise than if it was playing normal games and no throttling, thanks to that brilliant cooler. It did however exhibit that annoying revving problem I've posted about which is a design fault.
For this reason alone, I don't recommend this card like I used to, despite its improved framerate performance over a GTX 1080 FE and even some other non-reference cards and cooler temperatures. Idiotic problem with a simple and cheap solution that Zotac should have implemented.
I've benched them using Unigine Heaven 4.0 and the tl;dr version is that the GT 1030 has about 55% of the performance of the GTX 580 in DX9 and DX11 modes. Not too bad for such a low end card, even with the large generational gap between them.
Now, with that out of the way, I thought I'd have a bit of a benchathon this weekend, so expanded the number of cards to test, all of which I own. I have several more cards than this, but there has to be a limit somewhere, lol.
Here are the contenders, with TPU review links where possible. The review may not be of the same exact model, but will be functionally identical.
Zotac GTX 1080 Amp! Extreme Edition 8GB
MSI GT 1030 OC Edition 2GB (with passive cooling)
MSI GTX 780 Ti Gaming 3GB
Zotac GTX 580 Amp! Edition 1.5GB (reference cooler)
Palit GT 520 2GB (with active cooling)
Zotac GTX 285 1GB (reference cooler)
XFX 8800 GTX 768MB (reference cooler)
Just a note about the older cards not working. I had trouble getting the much older 8800 GTX and GTX 285 to actually give a picture, although the PC gave the POST beep and Windows booted otherwise. Turns out the gold PCIe connector pads had developed some visible tarnish on them which prevented good contact being made. Cleaning them with isopropyl alcohol fixed the problem, with the cards working perfectly after that. One to watch out for if your card plays up.
DRIVER
The oldest cards, the 8800 GTX and GTX 285 are no longer fully supported by NVIDIA so were run with the latest available Windows 10 64-bit driver, version 342.01. All the rest were run with the very latest version, 384.76.
The test rig is the one in my current system specs:
![](/forums/proxy.php?image=http%3A%2F%2Fimg.techpowerup.org%2F170716%2Fsystem-specs.png&hash=d25b39530992218072a3c2fec69058e5)
I used Unigine Heaven 4.0 in DX11 and DX9 modes. Since the 8800 GTX and GTX 285 only support DX10, I had to drop down the DX version for it to work. I chose DX9, because that was widely used, while hardly anything was released using DX10 and hence not representative of real world performance.
I used the Ultra quality settings, as shown:
DX11 with Normal tessellation level
![](/forums/proxy.php?image=http%3A%2F%2Fimg.techpowerup.org%2F170716%2Funigine-4-0-dx11.png&hash=7fddd2fce371a0dba61e1cb70d987ab0)
DX9
![](/forums/proxy.php?image=http%3A%2F%2Fimg.techpowerup.org%2F170716%2Funigine-4-0-dx09.png&hash=7a31f2bec4e8925cf25fc8e92732c05a)
RESULTS
Note that it wasn't possible to run the tests with ambient occlusion switched off for some reason. I would have preferred not to use it, because in my opinion it doesn't add that much to picture quality while hammering performance, so it's not worth it. This was especially true with the weaker cards and made them look a lot worse than they were otherwise. Dropping the quality down to Medium helped a lot, too.
![](/forums/proxy.php?image=http%3A%2F%2Fimg.techpowerup.org%2F170716%2Fdx11-bar-chart.png&hash=3fe01500f2241e2d3c802036e39bc78c)
![](/forums/proxy.php?image=http%3A%2F%2Fimg.techpowerup.org%2F170716%2Fdx09-bar-chart.png&hash=37cf9e4723acf9cb0c3f363e9f5f5d2a)
One thing to strike me immediately, is that all the DX11 cards gave better performance in DX11 mode, even with tessellation on. With tessellation off the performance was considerably better, putting it significantly ahead of DX9. This is in line with the efficiency improvements touted for DX11.
Unsurprisingly, the GTX 1080 totally dominated all the other cards, with the GTX 780 Ti coming in a distant second place, having about 62% of the performance. This is still enough for good performance in many games where the 3GB RAM limitation isn't a problem.
The GTX 580 made a reasonable job of smoothly rendering the scenes at these punishing settings, with many of them being rendered a lot faster and hence smoothly than the average 45-50fps seen here would suggest. It did it with a lot of huffing and puffing though with that noisy reference cooler! Killing ambient occlusion would have improved the performance considerably, resulting in perfectly smooth animation throughout.
The GT 1030 gave about 55% of this performance with the animation looking decidedly stuttery, but just about playable at a push. Again, just killing that pesky ambient occlusion would have resulted in markedly better performance, offering reasonably smooth framerates at about double the speed. Putting the quality setting to Medium would then result in very respectable performance indeed.
So, if you have a GTX 580 don't replace it with a GT 1030 expecting better framerates. Only do so maybe if you're not playing demanding games and want a quieter card that consumes much less power and has newer features. Even then, at £65 to £70 it's not exactly cheap right now, so you're probably better off going with a GT 1050 if you can, which offers much better performance.
It's interesting to see how the legendary 8800 GTX from waay back in 2006 was unplayable at these settings although it was a DX9 powerhouse back then. It even supported the brand new DX10, for which there was next to nothing that used it at the time. Shows just how far the mighty have fallen! Even without ambient occlusion, it wasn't really up to much. Hit it with some older games though and it will still rock.
The GTX 285 was about 50% better, but at these framerates still useless. Back in its day it felt like a supercharged 8800 GTX.
And now we arrive at the runt of the litter: the little GT 520. At roughly 2.6fps, it was a full-on slideshow and so much so that the application's interface was hammered too, with me barely able to select the options. Turning off ambient occlusion gave all of 4-5fps. Wow. This means that its big brother of the same generation, the GTX 580 is about 20 times faster, a staggering difference! And the GTX 1080 completely embarrasses it with 70 times the performance.
Comparing the two bottom end cards, the GT 1030 and the GT 520 showed just how far low end cards have come, with the GT 1030 having about 11 times the performance. For anyone thinking that these low end cards have stagnated, think again - Pascal is potent even in its smallest form. Nowadays they have to beat half decent integrated graphics processors, so they can't be too low performance.
NOISE PERFORMANCE
Of course, another important parameter is the noise that these cards make when running a stress test like this and boy were some of them loud!
The 8800 GTX and GTX 580 where by far the noisiest, being really obvious about the effort they were making, with the 8800 GTX sounding like it was going to take off. The GTX 285 was still pretty loud, but better. All of those cards were still reasonably tolerable however, because there was just a lot of windrush noise rather than any irritating buzzes or whines. Credit to NVIDIA for thinking about this aspect of performance. Note that they all sound a lot quieter when normal gaming.
Surprisingly, the GT 520 was pretty quiet, possibly because it may not have a temperature controlled fan and the GPU was cooking...
The quietest card was unsurprisingly the passive GT 1030, which didn't even exhibit any perceptable coil whine and worked with total stability under heavy stress. Impressive. I wouldn't risk Furmark on it though with that passive cooler. It was pretty hot to the touch afterwards though, as were the rest of the cards.
The GTX 780 Ti made a clearly audible noise, but still pretty good for a card being stressed to the max, showing what a good non-reference cooler can do. I can tell you that in normal gaming this card is very quiet indeed.
Finally, the GTX 1080 was whisper quiet, taking all this in its stride, not making any more noise than if it was playing normal games and no throttling, thanks to that brilliant cooler. It did however exhibit that annoying revving problem I've posted about which is a design fault.
For this reason alone, I don't recommend this card like I used to, despite its improved framerate performance over a GTX 1080 FE and even some other non-reference cards and cooler temperatures. Idiotic problem with a simple and cheap solution that Zotac should have implemented.
Last edited: