Reinventing Sigil
Yesterday (one day before they had promised) my components arrived. I spent a good portion of the night attempting to get the heatsink on right, much less having everything work, but at about 1 AM I looked at my windows desktop for the first time.
So when I started the computer had this:
* MSI K7D Master-L (slightly broken)
* 2x AMD Athlon MP 2400+ (operating at 1.5 GHz)
* 1x Generic 1 Gig stick of 2400 DDR (clocked at 100 GHz because the motherboard is broken)
* Asus nVidia Geforce Ti 4200 8x (running at 2x because of the broken board)
And then here's what I bought to replace those parts:
* Asus P5WD2
* Intel Pentium 4 640
* 2x Corsair ValueRAM 512 MB (800 MHz)
* MSI nVidia Geforce 7800 GT
It took a while just to get it all unpacked and get a good look at it. It's amazing how much comes with four FRUs. I wasn't prepared, either, because they had said it would arrive on Saturday, so I needed to get the stuff off of my computer that I wanted in preparation for a clean wipe of the hard drives.
A lot of stuff went wrong in terms of the software. And then a lot of stuff appeared to be wrong with the hardware, but it wasn't really wrong, I'm just a weakling. I had to really bash the CPU heatsink support pins to get them to stay in. I wonder if a lot of weak nerds have this problem?
Anyway, once I had all of the parts in place and the operating system installed I started benchmarking.
At this point, let me back up and give a little bit of back story. I really wanted to be able to give a before and after shot of my computer, but just taking a picture doesn't even come close to describing the changes, or course. So I benchmarked my computer under the old setup first, and then afterward under the new setup, and then posting the percentage of change.
BTW, I've done stuff similar to this before.
I was talking to my friend Matt at Cal Poly about this, and we started placing bets. Not all of the benchmarks would come out the same, obviously, since, for example, the hard drives are all exactly the same as they were before, so PCMark04's HDD test should be identical, or near identical. At this point I had a full list of the stats I had taken down for the old setup, and we talked numbers.
He said I'd average at a 500% improvement. I said 800%. And then I joked about having a 1000% improvement. He scoffed at that. I don't think he properly understood how bad my computer is.
Anyway, here is the original set of benchmarks:
The Old Setup
Keep in mind with all of these statistics that higher is better. On the artificial benchmarks I've underlined the most important statistic, or the one that should be most indicative of overall performance/improvement. On the two games, I underlined the quality setting at which I would like to run, and, therefor, the most important of the settings. So these five statistics are the ones I will be using to determine improvement.
And then I spent all of today benchmarking. The results, in a lot of ways, are really better than I could have hoped for. Take a look for yourself:
The New Setup
So the critical statistics changed as follows:
3DMark03: 1016%
YUS! This is exactly what I've been looking forward to. I couldn't use 05 unfortunately, because the Ti4200 doesn't support DX9, so it wouldn't even run on the old setup. But what can I say, this is stellar! If I'm forced to attribute it mostly to any one factor, I'd say it's the increase in memory frequency (going from 100 MHz to 800 GHz in dual channel is quit a big jump.)
PCMark04: 159%
This is the lowest improvement. My only explanation for this is that things like browsing the Internet and word processing really haven't changed in 3 years; there's really no reason for any of this to change, it was just fine before. It seems odd to me that the memory changed so little, and that the HDD rating changed so much, considering that the memory was so different and the hard drives were the same. Another one of life's great mysteries.
Aquamark3: 435%
It's amazing watching this program run. It now only takes about 45 seconds for the benchmark to go from start to finish. It sure zips. Decent improvement overall, well worth $900, I'll live with it.
Doom3: 1141%
*grins* Yeah. I figured an nVidia card would eat this game alive. It's especially funny considering I could only just barely push it below 80 FPS when, in any situation other than benchmarking, there is an imposed 60 framescap. Meaning, I should barely be able to tell the difference between 60 FPS and 90 FPS. Needless to say, it runs fast. Also, looking at the chart, I recognize the old pattern of a bottleneck. The whole upper left half of the chart moves only a few FPS. I think it's probably the processor (Intels are not the best at games) but I'm not sure. More testing will tell.
Farcry: 1148%
It's amazing to see how close this statistic is to the Doom III one. I guess it means that they were both similarly incapacitated by the old system. That demo is ugly, but a lot of people use it so I wanted to be able to compare my scores with theirs. There was a strange error on the Max and Ultra settings that caused many of the surfaces to look washed out on the new system, but it mostly looked pretty good. Again, more testing to know. Also, it seems strange to me that resolution should play such a small part in the performance.
I haven't yet run the anything newer that would require DX9 because I'm not really well settled down in the new installation. I had to pull everything out today when I found a set of mounting screws I'd been looking for. And right now I'm installing a clean copy of WXP after having wiped the HD. For the first time since I've owned this computer I'm finally routing the internal USB ports onto the front panel. I'll get some use out of that feature of my case I'd never used before. I'm spending a really large amount of time making sure that all this is right the first time around.
There is one part of all this that is not nearly finished yet. Unfortunately it looks like the whole system needs a slightly more beefy PSU, so I might be buying that soon. Yay, more expenses. The replacement of my old one won't be too difficult, though. I just don't want to be too satisfied too early for fear that something is wrong that I'm not noticing, like the last time when it took me over a year to realize that my FSB was totally gimped, and by that time of course I couldn't simply RMA. I want to be sure this time around.
So if the goal of the whole thing could be boiled down to one issue, I would say it's, "Can I play Elder Scrolls IV: Oblivion on it?" And I'd say, the way it looks so far, "Yes, with all of the settings turned all the way up."
So when I started the computer had this:
* MSI K7D Master-L (slightly broken)
* 2x AMD Athlon MP 2400+ (operating at 1.5 GHz)
* 1x Generic 1 Gig stick of 2400 DDR (clocked at 100 GHz because the motherboard is broken)
* Asus nVidia Geforce Ti 4200 8x (running at 2x because of the broken board)
And then here's what I bought to replace those parts:
* Asus P5WD2
* Intel Pentium 4 640
* 2x Corsair ValueRAM 512 MB (800 MHz)
* MSI nVidia Geforce 7800 GT
It took a while just to get it all unpacked and get a good look at it. It's amazing how much comes with four FRUs. I wasn't prepared, either, because they had said it would arrive on Saturday, so I needed to get the stuff off of my computer that I wanted in preparation for a clean wipe of the hard drives.
A lot of stuff went wrong in terms of the software. And then a lot of stuff appeared to be wrong with the hardware, but it wasn't really wrong, I'm just a weakling. I had to really bash the CPU heatsink support pins to get them to stay in. I wonder if a lot of weak nerds have this problem?
Anyway, once I had all of the parts in place and the operating system installed I started benchmarking.
At this point, let me back up and give a little bit of back story. I really wanted to be able to give a before and after shot of my computer, but just taking a picture doesn't even come close to describing the changes, or course. So I benchmarked my computer under the old setup first, and then afterward under the new setup, and then posting the percentage of change.
BTW, I've done stuff similar to this before.
I was talking to my friend Matt at Cal Poly about this, and we started placing bets. Not all of the benchmarks would come out the same, obviously, since, for example, the hard drives are all exactly the same as they were before, so PCMark04's HDD test should be identical, or near identical. At this point I had a full list of the stats I had taken down for the old setup, and we talked numbers.
He said I'd average at a 500% improvement. I said 800%. And then I joked about having a 1000% improvement. He scoffed at that. I don't think he properly understood how bad my computer is.
Anyway, here is the original set of benchmarks:
The Old Setup
Keep in mind with all of these statistics that higher is better. On the artificial benchmarks I've underlined the most important statistic, or the one that should be most indicative of overall performance/improvement. On the two games, I underlined the quality setting at which I would like to run, and, therefor, the most important of the settings. So these five statistics are the ones I will be using to determine improvement.
And then I spent all of today benchmarking. The results, in a lot of ways, are really better than I could have hoped for. Take a look for yourself:
The New Setup
So the critical statistics changed as follows:
3DMark03: 1016%
YUS! This is exactly what I've been looking forward to. I couldn't use 05 unfortunately, because the Ti4200 doesn't support DX9, so it wouldn't even run on the old setup. But what can I say, this is stellar! If I'm forced to attribute it mostly to any one factor, I'd say it's the increase in memory frequency (going from 100 MHz to 800 GHz in dual channel is quit a big jump.)
PCMark04: 159%
This is the lowest improvement. My only explanation for this is that things like browsing the Internet and word processing really haven't changed in 3 years; there's really no reason for any of this to change, it was just fine before. It seems odd to me that the memory changed so little, and that the HDD rating changed so much, considering that the memory was so different and the hard drives were the same. Another one of life's great mysteries.
Aquamark3: 435%
It's amazing watching this program run. It now only takes about 45 seconds for the benchmark to go from start to finish. It sure zips. Decent improvement overall, well worth $900, I'll live with it.
Doom3: 1141%
*grins* Yeah. I figured an nVidia card would eat this game alive. It's especially funny considering I could only just barely push it below 80 FPS when, in any situation other than benchmarking, there is an imposed 60 framescap. Meaning, I should barely be able to tell the difference between 60 FPS and 90 FPS. Needless to say, it runs fast. Also, looking at the chart, I recognize the old pattern of a bottleneck. The whole upper left half of the chart moves only a few FPS. I think it's probably the processor (Intels are not the best at games) but I'm not sure. More testing will tell.
Farcry: 1148%
It's amazing to see how close this statistic is to the Doom III one. I guess it means that they were both similarly incapacitated by the old system. That demo is ugly, but a lot of people use it so I wanted to be able to compare my scores with theirs. There was a strange error on the Max and Ultra settings that caused many of the surfaces to look washed out on the new system, but it mostly looked pretty good. Again, more testing to know. Also, it seems strange to me that resolution should play such a small part in the performance.
I haven't yet run the anything newer that would require DX9 because I'm not really well settled down in the new installation. I had to pull everything out today when I found a set of mounting screws I'd been looking for. And right now I'm installing a clean copy of WXP after having wiped the HD. For the first time since I've owned this computer I'm finally routing the internal USB ports onto the front panel. I'll get some use out of that feature of my case I'd never used before. I'm spending a really large amount of time making sure that all this is right the first time around.
There is one part of all this that is not nearly finished yet. Unfortunately it looks like the whole system needs a slightly more beefy PSU, so I might be buying that soon. Yay, more expenses. The replacement of my old one won't be too difficult, though. I just don't want to be too satisfied too early for fear that something is wrong that I'm not noticing, like the last time when it took me over a year to realize that my FSB was totally gimped, and by that time of course I couldn't simply RMA. I want to be sure this time around.
So if the goal of the whole thing could be boiled down to one issue, I would say it's, "Can I play Elder Scrolls IV: Oblivion on it?" And I'd say, the way it looks so far, "Yes, with all of the settings turned all the way up."