Last week, I got the Corsair F120 SSD drive and thought I'd try it out in my PS3 to see how much of a benefit they'd give.
I've already used this drive before successfully in a Linux server with quite impressive results. Running ext4, I've successfully had (sequential) read and write speeds of 250MB/sec.
These benchmarks aren't all that scientific, just a stopwatch and as a game benchmark, I used Gran Turismo 5 since it allows a big chunk of the game to be copied onto the harddrive and when doing so, doesn't use the bluray drive anymore. I didn't test how long the installation procedure took (it's definately a bit quicker than using the original SATA drive) but given that the bottleneck is probably mostly the bluray drive, I didn't bother.
Because the SSD harddrive was empty after mounting it to the PS3, I couldn't test the load times from the career mode. It probably wouldn't make much sense anyway, as the complexity would probably decrease the accuracy of the results. Instead, all tests I did was from the arcade mode and covers the loading time from the car selection screen to the race menu. Here they are:
I wasn't sure how much installing GT5 does even after you install the 8 GB of data, so as a precaution, I loaded each test-case twice. To eliminate another factor, I also made sure I wasn't logged onto PSN in both cases.
For the record, I built the SSD into my old fat PS3 and used my PS3slim with the original SATA drive for the test case. Both PS3s had the latest firmware on them. The reason I used this setup was because it was easier to compare them side by side and apart from the differing bluray drives, I'm not aware of any large changes to the hardware that should have an impact on load times. Later, testing the fat PS3 with its original SATA drive confirmed that the above load times are +/- equal to the times I did with the PS3slim.
My thoughts on the results are a bit mixed. On one hand, it shows how long the GT5 load times are despite the huge HDD install option. It's nice to see that the SSD can have such a positive benefit. On the other hand, factor 2 in most cases is a little disappointing given the Corsair can do up to 250MB (sequential read/write) and beyond in a normal Linux environment - with more than one thread easily between 65 and 130 MB/sec constantly. A "2.5 SATA drive in my experience on the same setup gets around 50-70MB/sec (sequential write) and with more than one thread not over 20MB/sec, not to mention slow latency. I'm also not sure why the chamonix test-case didn't have the usual factor 2 benefit, but those numbers were double checked. Perhaps the chamonix track is compressed higher.
The mediatomb test was a simple copy the file over the local 1Gbps lan network. Both have similar times with nothing really in it, so I suppose the limiting factor was some other component (perhaps network). 8 MB/sec is quite slow.
Is it worth it? Not for me - not for that price. Perhaps when SSDs become a bit cheaper, I may get one cheap one day for the PS3. I'm also not sure how optimized the PS3 filesystem is for SSD usage and how long such a drive would live. No way really to benchmark this sadly.
Feel free to add anything to the topic if you've done something similar or have questions. I made this topic specific to the PS3, but if it can be done on the Xbox (without getting it banned for using 3rd party hdds ) mods can feel free to change the title.
Cheers Phil
I've already used this drive before successfully in a Linux server with quite impressive results. Running ext4, I've successfully had (sequential) read and write speeds of 250MB/sec.
These benchmarks aren't all that scientific, just a stopwatch and as a game benchmark, I used Gran Turismo 5 since it allows a big chunk of the game to be copied onto the harddrive and when doing so, doesn't use the bluray drive anymore. I didn't test how long the installation procedure took (it's definately a bit quicker than using the original SATA drive) but given that the bottleneck is probably mostly the bluray drive, I didn't bother.
Because the SSD harddrive was empty after mounting it to the PS3, I couldn't test the load times from the career mode. It probably wouldn't make much sense anyway, as the complexity would probably decrease the accuracy of the results. Instead, all tests I did was from the arcade mode and covers the loading time from the car selection screen to the race menu. Here they are:
Code:
Gran Turismo 5 (v.1.03) / Arcade Professional A mode
Track Car Corsair F120 SSD Original SATA
-------------------------------------------------------------------------------------
London Zonda R 19.26s 39.47s
Nürburgring GP/F Zonda R 25.12s 46.45s
Capering Zonda R 20.03s 40.71s
Suzuka Zonda R 21.73s 43.82s
Chamonix Main changable Weather Ford RS WRC 21.25s 31.09s
Alternative Tests
Test Corsair F120 SSD Original SATA
-------------------------------------------------------------------------------------
Loading GT5 29s 55s
Copying 1566MB from Mediatomb Upnp to PS3 3.11m 3.12m
I wasn't sure how much installing GT5 does even after you install the 8 GB of data, so as a precaution, I loaded each test-case twice. To eliminate another factor, I also made sure I wasn't logged onto PSN in both cases.
For the record, I built the SSD into my old fat PS3 and used my PS3slim with the original SATA drive for the test case. Both PS3s had the latest firmware on them. The reason I used this setup was because it was easier to compare them side by side and apart from the differing bluray drives, I'm not aware of any large changes to the hardware that should have an impact on load times. Later, testing the fat PS3 with its original SATA drive confirmed that the above load times are +/- equal to the times I did with the PS3slim.
My thoughts on the results are a bit mixed. On one hand, it shows how long the GT5 load times are despite the huge HDD install option. It's nice to see that the SSD can have such a positive benefit. On the other hand, factor 2 in most cases is a little disappointing given the Corsair can do up to 250MB (sequential read/write) and beyond in a normal Linux environment - with more than one thread easily between 65 and 130 MB/sec constantly. A "2.5 SATA drive in my experience on the same setup gets around 50-70MB/sec (sequential write) and with more than one thread not over 20MB/sec, not to mention slow latency. I'm also not sure why the chamonix test-case didn't have the usual factor 2 benefit, but those numbers were double checked. Perhaps the chamonix track is compressed higher.
The mediatomb test was a simple copy the file over the local 1Gbps lan network. Both have similar times with nothing really in it, so I suppose the limiting factor was some other component (perhaps network). 8 MB/sec is quite slow.
Is it worth it? Not for me - not for that price. Perhaps when SSDs become a bit cheaper, I may get one cheap one day for the PS3. I'm also not sure how optimized the PS3 filesystem is for SSD usage and how long such a drive would live. No way really to benchmark this sadly.
Feel free to add anything to the topic if you've done something similar or have questions. I made this topic specific to the PS3, but if it can be done on the Xbox (without getting it banned for using 3rd party hdds ) mods can feel free to change the title.
Cheers Phil