Greetings,
I recently received a Sapphire Radeon X800 GT 256MB video card, and after using it for about a week, I will present my impressions to you in this mini-review.
Sapphire Tech website:
http://www.sapphiretech.com/
First off, the physical packaging and card itself. The box it came in I think was pretty standard. It included a DVI to VGA dongle, an S-Video cable, an S-Video to Component Video Cable, an S-Video to Composite Video Dongle, and a Composite Video Cable. Also included was a general "Getting Started" guide. On the back of the card are the standard ports: 1 VGA, 1 DVI, and 1 S-Video. The card itself has a blue PCB along with a small HSF and it does not require a dedicated connection to the power supply but instead draws all its power through the PCI Express x16 bus. About the fan...At start-up it is very noisy for a few seconds, and with the side of the case off it is very audible. However, unless the card is running under absolute full load and the fan is running at 100%, then with the case on, it's not so bad. However, if the card starts to heat up, it definitely noticeable by the fan, and the fan has a sort of "wheee"-ing noise, unlike my case and CPU fan, which has more of a "whooo"-ing sound. Anyway, on to the peculiarities of the X800 GT series on a whole...
There is much confusion about this card, mostly do to ATI's crazy X800 naming schemes and product cycles, but the X800 GT/O/^2 series was created to sell off ATI's old R423, R430, and R480 cores to be replaced by R5xx. For the X800 GT in particular, it can include either the R423 core or the R480 core. The Sapphire X800 GT 256MB includes the R423 core. If you're curious, the PowerColor X800 GT 256MB has the R480 core. The whole core business is of some importance in overclocking...
The card comes clocked default at 475/490 core/memory for the 256MB version, which includes GDDR3 memory. Just to let you know, the 128MB version comes clocked at only 475/350 core/memory and includes regular old DDR1 memory, so for $160 for the 256MB version and $130 for the 128MB version, I believe the 256MB version is a much better deal. Both cards come with 256-bit memory interfaces. Based on these specs and prices, it's not hard to guess the X800 GT is meant as a direct competitor to the classic GeForce 6600 GT. The X800 GT has twice the memory and twice the memory bandwidth of the 6600 GT, but the 6600 GT comes clocked default at 500/500 core/memory.
On the matter of overclocking, Sapphire has included overclocking software of their own called "TRIXX". It's a pretty useful program. It lets you select from a few different modes of overclocking by simply pressing a button: Standard (the card's stock values), Performance (the card's max values at the default fan speed, which seems to be 54% for the X800 GT 256MB), Silent (which runs the fan at 20% and reduces the clocks 30%), Extreme (fan running at 100% with max detected clocks), and Custom (which lets you set your own values for the core, memory, and fan speed). With the case off, I initially tried Performance, which gave me very stable values of 524.25/542.25 core/memory with no artifacts. However, when I tried to run this with the case on (partly due to my case's rather poor cooling with only one 120mm fan in the rear of the case), I got many artifacts and had to change the values of the core and memory. Instead of simply clocking the card down more, I decided to try to run the Extreme overclocking setting. This gave me 560.02/515.27 core/memory at 100% fan, but also gave me many artifacts and even crashed my system, so I definitely could not run at those speeds with the case on. After rebooting, I decided to try Performance again, but this time with the case on. This gave me 527.45/515.08 core/memory. With these values, ATITool did not detect any artifacts, but while playing the Battlefield 2 demo, there were slight artifacts when taking off and flying aircraft, and maybe in other less conspicuous areas, so I decided to try different settings. This time, I used ATITool to manually overclock the card, since I could scan for artifacts while overclocking. I stopped once I got the card to 525/525 core/memory which ran the fan at 67% during a 3D application, and at 54% (default) when in normal 2D mode. ATITool again did not report any artifacts at these values, but I still saw the same kind of artifacting in the Battlefield 2 demo. Again I went back to ATITool. This time I ran the card at 499.50/499.50 core/memory, or 500/500: the 6600 GT's clock speeds. I decided to stay here at least for now since I did not get any artifacts at these speeds in any applications, but I could push it higher slowly until I found a happy medium between 500/500 and 525/525. Back to the Sapphire TRIXX overclocking software, it also includes some other functions. It measures VPU and board temperatures, allows you to save different overclocking values for different game profiles, let's you monitor the VPU temperature in a graph, and of course lets you check for new version of the TRIXX software on the Internet, but also for new ATI Catalyst drivers on the Internet, in case you are too lazy to go to ati.com.
Oh yes, about the core. For anyone who is worried about having an R423 core instead of an R480 core, stop worrying. The cores are virtually identical. While the R480 can generally clock slightly higher than the R423, the R423 generally runs slightly cooler and consumes a few less watts power. I'm not too sure about unlocking extra pipelines, since all X800 GT's come with 8 pipelines like the 6600 GT, though the R423 and R480 cores both natively have 16 pipelines. However, as far as I know, I am under the impression that the extra 8 pipelines are hard-locked or laser cut and cannot be unlocked no matter what software or BIOS flash is used. Therefore, in order to ensure I do not corrupt the BIOS on my card or anything like that, I did not try to unlock the extra 8 pipelines.
Before I get to my performance in games, I'll talk a little more about the software bundle included with the card. It of course came with Catalyst drivers (Version 5.7) and the TRIXX software, but it's highly recommended that you just download the newest Catalyst drivers from ati.com (Version 5.10 as of this writing). The card also came with CyberLink PowerDVD 6, which is pretty useful for me since I needed DVD playback software on my PC anyway. It also comes with the full version of Tom Clancy's Splinter Cell: Pandora Tomorrow, which is the second in the series and is from 2004. Although I personally do not like the game, I'm sure some people will appreciate it being included with the card. Oh yeah, and it also came with one of those stickers you can put on your case to show you have a Sapphire/ATI card in your system, if you really want to, like you can with those Intel or AMD stickers and those Windows stickers, for example.
So now, on to the "benchmarking". I say that in quotes because it's not really benchmarking, since I did not use any real tools that could have told me average FPS or similar on a standard timedemo, but rather I'll tell you what I basically detected and found to be the average through console commands in individual games and the like. My basic system specs are as follows, and you can find more detailed specs in the link in my signature:
Athlon 64 3000+ Venice 1.8 GHz
1GB Dual Channel Corsair Value Select DDR RAM
MSI RS480M2-IL Motherboard
Antec SLK2650-BQE case with included Antec 350 Watt SmartPower2.0 power supply
Seagate 120GB IDE ATA/100 Hard Drive
NEC ND-3540A DVD Burner
Samsung 730B 17" LCD Monitor (1280 x 1024 native resolution, 8ms response time)
Sapphire Radeon X800 GT 256MB Video Card
By the way, my motherboard has a 24-pin main power supply connector, but my power supply's main power supply to the motherboard connector is the 20-pin version. However, my motherboard's manual says it is okay to connect a power supply like this to the motherboard, so mine is connected with the 20-pin connector with 4 of the pins unused.
Okay, now for the benchmarks. On a side note, before this video card I was using my Radeon Xpress 200 RS480 IGP, so I'll try to give some FPS numbers using that solution to compare to the ones I get using the X800 GT. Oh, and one more thing. Most of these tests were done on demos of games.
***
Battlefield 2 Demo
***
RS480
---
800 x 600 default (low) settings:
20 FPS average
Sapphire X800 GT 256MB
---
800 x 600 default (low) settings:
>60 FPS average, usually 80-90 FPS
1280 x 960 all settings highest with no AA (preferred settings) on a 64 player map:
Default clocks: 20 FPS average
500/500 core/memory: 30 FPS average
525/525 core/memory: 40 FPS average (slight artifacting)
***
Unreal Tournament 2004 Demo
***
RS480
---
1024 x 768 medium settings:
30 FPS average
Sapphire X800 GT 256MB
---
1280 x 1024 highest settings:
Default clocks: 60 FPS average
***
Call of Duty 2 Demo
***
RS480
---
Not even worth mentioning here. I had to force DirectX 7 mode just to get it close to playable since the absolute lowest settings and resolution in DirectX 9 mode produced unplayable framerates.
Sapphire X800 GT 256MB
---
I did not know how to read the FPS in this game, so I will say what it looked like I was getting.
1280 x 1024 default (medium) settings:
20 FPS average in much action to 30 FPS average with less action
***
Tom Clancy's Splinter Cell: Pandora Tomorrow
***
I did not test this game on my RS480 IGP, since I installed the game after I installed the new video card.
Sapphire X800 GT 256MB
---
1280 x 1024 highest settings:
60 FPS average, with the occasional minimum being 30 FPS
***
Final Fantasy XI Official Benchmark 3
***
I did no record my results with my RS 480 IGP, but for the High Resolution mode of the benchmark, I think they were ~2000-H and I can't remember the results for the Low Resolution mode, unless ~2000-L was actually my results for the Low Resolution mode...T_T Anyway...
Sapphire X800 GT 256MB
---
Low Resolution: 5037-L
High Resolution: 3436-H
Interpretations for these results and the benchmarking program itself can be found at:
http://www.playonline.com/ff11us/multimedia/download/bench/
NOTE: I may be adding more games as I play them and more benchmarks periodically, along with the possibility of more accurate benchmarks and more information.
You can get a pretty good idea of the performance of this card from these numbers. I'm not going to spend much time writing a conclusion for this mini-review, so you can make your own conclusion. Anyway, for $160 I'd definitely recommend the Sapphire X800 GT 256MB video card if you can forgo Shader Model 3.0 support and "put up with" Shader Model 2.0b support. Of course, for $130, the 6600 GT with Shader Model 3.0 support can be obtained, but it has half as much memory and half the memory bandwidth of the X800 GT, so that may be a concern when deciding what to purchase in the low midrange ~$150 segment. To end this review, I'd like to say that, in the end, if you are in the market for a low midrange ~$150 video card, the your choice of purchase is ultimately up to you and I am only trying to provide you with more information about one of your choices, and I hope that you got something out of this review that will help to influence your ultimate purchase decision.
Thank you.
I recently received a Sapphire Radeon X800 GT 256MB video card, and after using it for about a week, I will present my impressions to you in this mini-review.
Sapphire Tech website:
http://www.sapphiretech.com/
First off, the physical packaging and card itself. The box it came in I think was pretty standard. It included a DVI to VGA dongle, an S-Video cable, an S-Video to Component Video Cable, an S-Video to Composite Video Dongle, and a Composite Video Cable. Also included was a general "Getting Started" guide. On the back of the card are the standard ports: 1 VGA, 1 DVI, and 1 S-Video. The card itself has a blue PCB along with a small HSF and it does not require a dedicated connection to the power supply but instead draws all its power through the PCI Express x16 bus. About the fan...At start-up it is very noisy for a few seconds, and with the side of the case off it is very audible. However, unless the card is running under absolute full load and the fan is running at 100%, then with the case on, it's not so bad. However, if the card starts to heat up, it definitely noticeable by the fan, and the fan has a sort of "wheee"-ing noise, unlike my case and CPU fan, which has more of a "whooo"-ing sound. Anyway, on to the peculiarities of the X800 GT series on a whole...
There is much confusion about this card, mostly do to ATI's crazy X800 naming schemes and product cycles, but the X800 GT/O/^2 series was created to sell off ATI's old R423, R430, and R480 cores to be replaced by R5xx. For the X800 GT in particular, it can include either the R423 core or the R480 core. The Sapphire X800 GT 256MB includes the R423 core. If you're curious, the PowerColor X800 GT 256MB has the R480 core. The whole core business is of some importance in overclocking...
The card comes clocked default at 475/490 core/memory for the 256MB version, which includes GDDR3 memory. Just to let you know, the 128MB version comes clocked at only 475/350 core/memory and includes regular old DDR1 memory, so for $160 for the 256MB version and $130 for the 128MB version, I believe the 256MB version is a much better deal. Both cards come with 256-bit memory interfaces. Based on these specs and prices, it's not hard to guess the X800 GT is meant as a direct competitor to the classic GeForce 6600 GT. The X800 GT has twice the memory and twice the memory bandwidth of the 6600 GT, but the 6600 GT comes clocked default at 500/500 core/memory.
On the matter of overclocking, Sapphire has included overclocking software of their own called "TRIXX". It's a pretty useful program. It lets you select from a few different modes of overclocking by simply pressing a button: Standard (the card's stock values), Performance (the card's max values at the default fan speed, which seems to be 54% for the X800 GT 256MB), Silent (which runs the fan at 20% and reduces the clocks 30%), Extreme (fan running at 100% with max detected clocks), and Custom (which lets you set your own values for the core, memory, and fan speed). With the case off, I initially tried Performance, which gave me very stable values of 524.25/542.25 core/memory with no artifacts. However, when I tried to run this with the case on (partly due to my case's rather poor cooling with only one 120mm fan in the rear of the case), I got many artifacts and had to change the values of the core and memory. Instead of simply clocking the card down more, I decided to try to run the Extreme overclocking setting. This gave me 560.02/515.27 core/memory at 100% fan, but also gave me many artifacts and even crashed my system, so I definitely could not run at those speeds with the case on. After rebooting, I decided to try Performance again, but this time with the case on. This gave me 527.45/515.08 core/memory. With these values, ATITool did not detect any artifacts, but while playing the Battlefield 2 demo, there were slight artifacts when taking off and flying aircraft, and maybe in other less conspicuous areas, so I decided to try different settings. This time, I used ATITool to manually overclock the card, since I could scan for artifacts while overclocking. I stopped once I got the card to 525/525 core/memory which ran the fan at 67% during a 3D application, and at 54% (default) when in normal 2D mode. ATITool again did not report any artifacts at these values, but I still saw the same kind of artifacting in the Battlefield 2 demo. Again I went back to ATITool. This time I ran the card at 499.50/499.50 core/memory, or 500/500: the 6600 GT's clock speeds. I decided to stay here at least for now since I did not get any artifacts at these speeds in any applications, but I could push it higher slowly until I found a happy medium between 500/500 and 525/525. Back to the Sapphire TRIXX overclocking software, it also includes some other functions. It measures VPU and board temperatures, allows you to save different overclocking values for different game profiles, let's you monitor the VPU temperature in a graph, and of course lets you check for new version of the TRIXX software on the Internet, but also for new ATI Catalyst drivers on the Internet, in case you are too lazy to go to ati.com.
Oh yes, about the core. For anyone who is worried about having an R423 core instead of an R480 core, stop worrying. The cores are virtually identical. While the R480 can generally clock slightly higher than the R423, the R423 generally runs slightly cooler and consumes a few less watts power. I'm not too sure about unlocking extra pipelines, since all X800 GT's come with 8 pipelines like the 6600 GT, though the R423 and R480 cores both natively have 16 pipelines. However, as far as I know, I am under the impression that the extra 8 pipelines are hard-locked or laser cut and cannot be unlocked no matter what software or BIOS flash is used. Therefore, in order to ensure I do not corrupt the BIOS on my card or anything like that, I did not try to unlock the extra 8 pipelines.
Before I get to my performance in games, I'll talk a little more about the software bundle included with the card. It of course came with Catalyst drivers (Version 5.7) and the TRIXX software, but it's highly recommended that you just download the newest Catalyst drivers from ati.com (Version 5.10 as of this writing). The card also came with CyberLink PowerDVD 6, which is pretty useful for me since I needed DVD playback software on my PC anyway.
It also comes with the full version of Tom Clancy's Splinter Cell: Pandora Tomorrow, which is the second in the series and is from 2004. Although I personally do not like the game, I'm sure some people will appreciate it being included with the card. Oh yeah, and it also came with one of those stickers you can put on your case to show you have a Sapphire/ATI card in your system, if you really want to, like you can with those Intel or AMD stickers and those Windows stickers, for example.So now, on to the "benchmarking". I say that in quotes because it's not really benchmarking, since I did not use any real tools that could have told me average FPS or similar on a standard timedemo, but rather I'll tell you what I basically detected and found to be the average through console commands in individual games and the like. My basic system specs are as follows, and you can find more detailed specs in the link in my signature:
Athlon 64 3000+ Venice 1.8 GHz
1GB Dual Channel Corsair Value Select DDR RAM
MSI RS480M2-IL Motherboard
Antec SLK2650-BQE case with included Antec 350 Watt SmartPower2.0 power supply
Seagate 120GB IDE ATA/100 Hard Drive
NEC ND-3540A DVD Burner
Samsung 730B 17" LCD Monitor (1280 x 1024 native resolution, 8ms response time)
Sapphire Radeon X800 GT 256MB Video Card
By the way, my motherboard has a 24-pin main power supply connector, but my power supply's main power supply to the motherboard connector is the 20-pin version. However, my motherboard's manual says it is okay to connect a power supply like this to the motherboard, so mine is connected with the 20-pin connector with 4 of the pins unused.
Okay, now for the benchmarks. On a side note, before this video card I was using my Radeon Xpress 200 RS480 IGP, so I'll try to give some FPS numbers using that solution to compare to the ones I get using the X800 GT. Oh, and one more thing. Most of these tests were done on demos of games.
***
Battlefield 2 Demo
***
RS480
---
800 x 600 default (low) settings:
20 FPS average
Sapphire X800 GT 256MB
---
800 x 600 default (low) settings:
>60 FPS average, usually 80-90 FPS
1280 x 960 all settings highest with no AA (preferred settings) on a 64 player map:
Default clocks: 20 FPS average
500/500 core/memory: 30 FPS average
525/525 core/memory: 40 FPS average (slight artifacting)
***
Unreal Tournament 2004 Demo
***
RS480
---
1024 x 768 medium settings:
30 FPS average
Sapphire X800 GT 256MB
---
1280 x 1024 highest settings:
Default clocks: 60 FPS average
***
Call of Duty 2 Demo
***
RS480
---
Not even worth mentioning here. I had to force DirectX 7 mode just to get it close to playable since the absolute lowest settings and resolution in DirectX 9 mode produced unplayable framerates.
Sapphire X800 GT 256MB
---
I did not know how to read the FPS in this game, so I will say what it looked like I was getting.
1280 x 1024 default (medium) settings:
20 FPS average in much action to 30 FPS average with less action
***
Tom Clancy's Splinter Cell: Pandora Tomorrow
***
I did not test this game on my RS480 IGP, since I installed the game after I installed the new video card.
Sapphire X800 GT 256MB
---
1280 x 1024 highest settings:
60 FPS average, with the occasional minimum being 30 FPS
***
Final Fantasy XI Official Benchmark 3
***
I did no record my results with my RS 480 IGP, but for the High Resolution mode of the benchmark, I think they were ~2000-H and I can't remember the results for the Low Resolution mode, unless ~2000-L was actually my results for the Low Resolution mode...T_T Anyway...
Sapphire X800 GT 256MB
---
Low Resolution: 5037-L
High Resolution: 3436-H
Interpretations for these results and the benchmarking program itself can be found at:
http://www.playonline.com/ff11us/multimedia/download/bench/
NOTE: I may be adding more games as I play them and more benchmarks periodically, along with the possibility of more accurate benchmarks and more information.
You can get a pretty good idea of the performance of this card from these numbers. I'm not going to spend much time writing a conclusion for this mini-review, so you can make your own conclusion. Anyway, for $160 I'd definitely recommend the Sapphire X800 GT 256MB video card if you can forgo Shader Model 3.0 support and "put up with" Shader Model 2.0b support. Of course, for $130, the 6600 GT with Shader Model 3.0 support can be obtained, but it has half as much memory and half the memory bandwidth of the X800 GT, so that may be a concern when deciding what to purchase in the low midrange ~$150 segment. To end this review, I'd like to say that, in the end, if you are in the market for a low midrange ~$150 video card, the your choice of purchase is ultimately up to you and I am only trying to provide you with more information about one of your choices, and I hope that you got something out of this review that will help to influence your ultimate purchase decision.
Thank you.
Facebook
Twitter