besweeet Posted November 15, 2010 Share Posted November 15, 2010 Check this out: http://img230.imageshack.us/img230/654/52032860.th.jpg I was shocked to see the difference in speeds... I should be getting 270MB/s read (doesn't happen, usually gets close to 200MB/s, which is fine IMO) and 130MB/s (was reached at the beginning and a month later). Any ideas? Link to comment Share on other sites More sharing options...
Synbios Posted November 15, 2010 Share Posted November 15, 2010 I would first provide system specs in your profile. You didn't even tell us what drive this is, but if you're expecting 270 MB/read then I am guessing it is a Force drive? I would suggest imaging the drive and then secure erase should bring back the speeds to the original state. Make sure AHCI is enabled as well as TRIM and I would use Windows 7 or another TRIM supporting OS with a SSD like the Force drive. If you're not using TRIM and you are filling this drive and making reads/writes on a daily basis then I would expect performance degradation like that. You should also check the alignment. Link to comment Share on other sites More sharing options...
besweeet Posted November 15, 2010 Author Share Posted November 15, 2010 Oops! I apologize. I have no idea why I forgot to post my specs. Anyway, I'm running a mid-2009 15" MacBook Pro. It has both a SSD and a 7200RPM HDD in it (I swapped the optical drive out with a little SATA caddy that's being used by the HDD). It's running Mac OS X 10.6.5 and Windows 7 Professional x64 (native via Boot Camp), both being booted via the SSD (split up into 2 partitions, both evenly spaced). Windows 7 has TRIM enabled, and AHCI should be enabled (via EFI's BIOS thing or whatever). I've already taken most of the precautions when it comes to keeping a SSD healthy (moving cache files/temp files/etc over to the HDD, etc). The 270MB/s read came from Newegg, but I don't think that their information is correct. I use Mac OS X most of the time, and it isn't fine-tuned for use on a SSD, but it seems to work great. Overall, the performance seems to be like it was on day one. I haven't noticed any performance decreases whatsoever, but those numbers are worrying me a bit. Link to comment Share on other sites More sharing options...
Synbios Posted November 15, 2010 Share Posted November 15, 2010 What is the actual SSD and I will tell you the actual advertised speeds. Your performance definitely degraded between october and november and this I'm afraid is due to running OSX as currently it does not support TRIM. Having TRIM on Windows 7 is not enough to make up for the fact that your other OS doesn't have it because EVERYTIME you delete or move a file while you're in OSX you're degrading the performance of your drive. I've already taken most of the precautions when it comes to keeping a SSD healthy (moving cache files/temp files/etc over to the HDD, etc). Moving files will make the situation worse actually, the only way to prevent the drive from being degraded is by not writing the files in the first place and ensuring that you save files to the HDD. Whenever you see the performance to degrade significantly your best bet is to image the drive, secure erase and then restore. edit: Let me also add that I just noticed that you've partitioned your SSD. This will also degrade performance since performance is relative to the amount of free space on the drive. Link to comment Share on other sites More sharing options...
besweeet Posted November 15, 2010 Author Share Posted November 15, 2010 Oops, I still didn't post it all :(. It's a Corsair V64. I didn't mean to say "move", but just change the temporary files directory and cache directories. I'm not seeing any decrease in performance, but the numbers say otherwise. I'm still very happy with the performance (when compared to using my HDD)... Link to comment Share on other sites More sharing options...
Synbios Posted November 15, 2010 Share Posted November 15, 2010 Yes even a degraded SSD should have much better performance than a HDD. http://www.corsair.com/products/ssd_nova/default.aspx Nova's speed is listed at 270 MB/sec for read and 195 MB/sec for write but I believe that is a typo that newegg has copied as well. The Nova 32 drives are rated at 195 MB/s sequential read and 70 MB/s sequential write so you are right on with the performance it looks to be running fine. Link to comment Share on other sites More sharing options...
Corsair Employees RAM GUY Posted November 16, 2010 Corsair Employees Share Posted November 16, 2010 Nova 64 SSD's are rated at 215 MB/s sequential read and 130 MB/s sequential write. I would try and image the current drive then use parted magic to secure erase the drive. What O.S. are you using and are you running a single drive? Link to comment Share on other sites More sharing options...
Wired Posted November 16, 2010 Share Posted November 16, 2010 As a clarification the speed is listed under the Resources tab where Synbios linked to. Ram Guy: Anyway, I'm running a mid-2009 15" MacBook Pro. It has both a SSD and a 7200RPM HDD in it (I swapped the optical drive out with a little SATA caddy that's being used by the HDD). It's running Mac OS X 10.6.5 and Windows 7 Professional x64 (native via Boot Camp), both being booted via the SSD (split up into 2 partitions, both evenly spaced). Windows 7 has TRIM enabled, and AHCI should be enabled (via EFI's BIOS thing or whatever). I've already taken most of the precautions when it comes to keeping a SSD healthy (moving cache files/temp files/etc over to the HDD, etc). The 270MB/s read came from Newegg, but I don't think that their information is correct. I use Mac OS X most of the time, and it isn't fine-tuned for use on a SSD, but it seems to work great. Overall, the performance seems to be like it was on day one. I haven't noticed any performance decreases whatsoever, but those numbers are worrying me a bit. Link to comment Share on other sites More sharing options...
Synbios Posted November 16, 2010 Share Posted November 16, 2010 As a clarification the speed is listed under the Resources tab where Synbios linked to. Oh thanks for pointing that out, I didn't see that before. That's a very good reference. Link to comment Share on other sites More sharing options...
Corsair Employees RAM GUY Posted November 16, 2010 Corsair Employees Share Posted November 16, 2010 If you have the SSD is the secondary position please move it to the primary S-ATA controller and re-test it. It is known the secondary port is limited to S-ATA1 on many systems including MAC's If that does not solve it I would make and image of the drive and use parted Magic to Secure Erase it and then quick format it in Windows 7 and test its performance as a second drive then delete the partition and image it back. Link to comment Share on other sites More sharing options...
besweeet Posted November 18, 2010 Author Share Posted November 18, 2010 If you have the SSD is the secondary position please move it to the primary S-ATA controller and re-test it. It is known the secondary port is limited to S-ATA1 on many systems including MAC's If that does not solve it I would make and image of the drive and use parted Magic to Secure Erase it and then quick format it in Windows 7 and test its performance as a second drive then delete the partition and image it back. That doesn't matter. The SSD has been in the same spot since day one. Oh well. Performance is still great. Link to comment Share on other sites More sharing options...
Synbios Posted November 18, 2010 Share Posted November 18, 2010 That doesn't matter. The SSD has been in the same spot since day one. Oh well. Performance is still great. It's still pretty obvious to me that your drive is slowing down because you use a Mac and they don't currently support TRIM. Even if you haven't noticed the decrease in speed yet, the numbers show it. Link to comment Share on other sites More sharing options...
besweeet Posted November 19, 2010 Author Share Posted November 19, 2010 Did another test today, and speeds were considerably slower than a few days ago: http://img831.imageshack.us/img831/5117/attobenchmark4.jpg Yet I see absolutely no decrease in performance... Oh well. I'm still satisfied. I'll just ignore those numbers. Link to comment Share on other sites More sharing options...
besweeet Posted December 21, 2010 Author Share Posted December 21, 2010 Just an update... With no change to my daily computing habits, I went ahead and did my monthly ATTO benchmark (default settings). Speeds are much higher, and are even higher than my initial benchmark months ago! 12/21/2010 (today): http://img84.imageshack.us/img84/5117/attobenchmark4.jpg 11/18/2010 (when the speeds took an unexpected dump): http://img90.imageshack.us/img90/4281/attobenchmark3.jpg 10/22/2010: http://img525.imageshack.us/img525/8994/attobenchmark2.png 9/20/2010 (day I purchased the SSD): http://img401.imageshack.us/img401/121/attobenchmark1.png Now, I will say that I did a few things earlier today. I wanted to see how long it took to open 44 different apps. So, I made a little batch file (here it is if you're wondering how I wrote it). I opened the 44 apps 3 times to test out how long it would take. Could putting a whole bunch of random read/writes onto the drive help it out like this? Thoughts would be appreciated. ALSO! For those who are wondering, here are my SSD Status results. 12/21/2010 (today): http://img35.imageshack.us/img35/946/ssdstatus4.png 11/18/2010: http://img809.imageshack.us/img809/5467/ssdstatus3.png 10/22/2010: http://img145.imageshack.us/img145/4489/ssdstatus2.png 9/20/2010: http://img716.imageshack.us/img716/6415/ssdstatus1.png Seems like everything is back to normal. Link to comment Share on other sites More sharing options...
Yellowbeard Posted December 21, 2010 Share Posted December 21, 2010 2 things of note for you. 1. Once an SSD gets up to about 85%+ full, it's going to slow down some. 2. Benchmarking writes a LOT of data to the drive which is also going to put more blocks into a written state and contribute to your decline in speeds. I'd minimize the benching if you want to maintain peak performance. Link to comment Share on other sites More sharing options...
besweeet Posted December 21, 2010 Author Share Posted December 21, 2010 2 things of note for you. 1. Once an SSD gets up to about 85%+ full, it's going to slow down some. 2. Benchmarking writes a LOT of data to the drive which is also going to put more blocks into a written state and contribute to your decline in speeds. I'd minimize the benching if you want to maintain peak performance. I already know those things. My drive is about 65% filled. I only do benchmarks once a month. Did you even see my screenshots? Link to comment Share on other sites More sharing options...
Yellowbeard Posted December 21, 2010 Share Posted December 21, 2010 I already know those things. My drive is about 65% filled. I only do benchmarks once a month. Did you even see my screenshots? I glanced at them. In my opinion, and especially on a drive 65% full, once per month is too frequent. But, it's your drive ::pirate:: Link to comment Share on other sites More sharing options...
besweeet Posted December 21, 2010 Author Share Posted December 21, 2010 I glanced at them. In my opinion, and especially on a drive 65% full, once per month is too frequent. But, it's your drive ::pirate:: I understand that, but I haven't lost any performance over the course of 3 months. Link to comment Share on other sites More sharing options...
Yellowbeard Posted December 21, 2010 Share Posted December 21, 2010 I understand that, but I haven't lost any performance over the course of 3 months. Pure performance is not the only thing to consider. Write amplification and P/E cycles wear your drive out faster. Frequent benching adds to this. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.