Unstable new build failing prime95 blend test

[caesparktom]

New Build is failing prime95 blend test. After 12 hours blend w/ 8 threads it crashes. It does not find any prime number errors and no bsod just shuts down and reboots to "windows did not shut down properly" screen.

fyi, I posted problem on my old new build thread here. Flong has been helping me troubleshoot but suggested I start a new thread. (Troubleshooting starts near end of page two.)

History/Steps taken so far:

Assembled new build (my first build) without major incident. Actually, I should mention that initially computer did failed to boot with a chassis intrusion error. I reset RTC RAM and didn't have any problems after that. So, I booted up. Installed win7 64 bit. Made no changes to bios and ran Prime95 blend test + Coretemp to monitor cpu temps. It crashed without errors after about 7 minutes--just shut down. I rebooted and confirmed that p95 results log was empty-no errors. Temperatures at full load during test were approx 53-57C.

Temperatures were a little higher than expected so I reseated cooler and re-ran blend test. This time if crashed after 5 minutes and temperatures were unchanged (53-57C).

I researched difference between blend, sml fft, and lrg fft tests here. Of interest was that fact that sml ffts test does not access memory as much as lrg ffts:

The "Small FFTs" test uses relatively small FFTs which can fit into the CPU cache. As a result, the small FFT test is the one which accesses your main memory the least but it still makes some memory accesses. Prime95 automatically creates a FFT size range which will fit into the L2 cache of your CPU.

The "In-place large FFTs" test uses relatively large FFTs which cannot fit into the CPU cache so this test accesses main memory a lot. It only accesses a relatively small amount of main memory because it runs the FFTs in-place so it accesses the same RAM over and over.

I realize that this is not best way to test memory but decided to run the sml fft test because both times a ran the blend test, the first test it ran and crashed on was incidentally a large prime. So I ran and it ran for an hour without errors or crashing. I stopped it and concluded that it was a memory problem.

I checked bios ram settings (timing, frequency and voltage) and found that they were off by a bit. So i set explicitly to manufacture specs. Re-ran blend test. This time it ran 13 hours, produced no errors, but did still crash... again no bsod just sudden crash. Here are some temp readings during first hour or so.

time | c1 | c2 | c3 | c4
5:00 | 48c | 51c | 55c | 52c
5:13 | 52c | 54c | 58c | 55c
6:50 | 53c | 53c | 57c | 55c
.
.
.
temps were pretty consistent for duration of test. Actually I turned A/C off for bit to see how temps reacted and core 1,2 and 3 went up to ~ 60c and core 3 to 64c. But, it IS pretty hot here in beijing without A/C. I should note that crash occurred in middle of night when A/C was off so temps were probable around 60-64c before crash. btw, should i be concerned that core3 is always about 4c hotter than other cores? btw, GPU temp also around 52C.

Anyways, Stock system w/no Overclock should have no trouble burning for 24+ hours right? Not sure what to do now. Seems my system is just not quite as stable as it should be and not cooling quite as well as i would like. What do you suggest? I would like overclock a bit, but if my stock system is already marginal in terms of stability i don't know that i have any room to push it.

Btw, this is a workstation build so stability is relatively important.

Specs:

MOBO: ASUS P8P67 WS REVOLUTION
CPU: i7 2600k
GPU: NVidia Quadro 4000
RAM: G.SKILL Ripjaws X Series 8GB (2 x 4GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) F3-12800CL9D-8GBXL --- x2(16GB)
SSD: OCZ vertex 3 （120GB）
HDD: SAMSUNG Spinpoint F3 HD103SJ 1TB ---x2(2TB)
DVD: Sony Optiarc AD-7240S
LCD: Dell UltraSharp U2410
Case: Corsair 650D
PSU: Corsair HX750
COOLER: Corsair H60

Steps I am considering:

1. replace ram w/ same ram or different ram(perhaps try different freq/timing)
2. upgrade cooler to H80 or H100.
3. jump out the window and take my new build with me

thanks for any help

 

[jaquith]

System Restore Example - say you update {anything} and Windows won't boot or boot properly. Choices 2 minute System Restore <OR> Re-Install the OS. I never recommend any SSD smaller than 120GB.

Yeah, sure people parrot what they read with no thinking of the consequences behind it. You can limit the System Restore size, and in Doug's post you'll notice my comments. Doug and I are friends and here once in a while disagreed. I don't baby my SSD's, and I don't feel like losing data over a few GB's nor sucking up a few hours of my time over a few GB's more. I run UPS on 100% of my systems, but you're screwed if Hibernation is off and they're unattended 'poof' goes hours of expensive $ labor.

edit: Once you see the difference between Hybrid Sleep and Standby in KWH you'll know to allow Hybrid Sleep {S3+S4}. Yes, if you have Hybrid Sleep Enabled there's a hiberfil.sys file on your SSD 😉

The PST {Personal Storage Table} 'POP' email and OST {Off-line Storage Table} 'Exchange' Outlook email files. Once their 'Indexed' they're pretty fast on either SSD or HDD. On our 'PCs' we have SSD {OS, Apps, & Working Data} and RAID 1 HDD {Documents, etc, and general archival Data}. Once setup those files and folders are 'transparent' once you 'move'/change the Default locations. Nice article, but REMEMBER to create FOLDERS and NOT use e.g. 'D' or top level drive for a 'folder'; see -> http://headstrongfarm.hubpages.com/hub/Windows-7_-_Moving_My_Documents

A SSD/HDD MB/s is only a fraction of the 'speed' - the latency 12ms vs 0.1ms and IOPS is the real measure of speed. The key advantages to SSD is 'data is data' wherever it's located on and SSD which is why there's no need to defragment an SSD and the next is time to access that data and on and SSD it's 120X+ faster and fragmentation doesn't matter on the SSD. There's more things but those are the biggies. The 4K test, ms access times, IOPS and BOTH Read/Writes are what draws my attention first, then the rest are the distance remainder from benchmarks. ATTO simply measures peak R/W under a variety of sizes. A good bench of many is AS SSD, but manufactures publish peak ATTO R/W.

Yes, both your an my HDDs are like dragging an anchor in the mud compared to any SSD. The SATA3 vs SATA2 ports is of little help on HDD(s), and if using an SSD I recommend leaving the HDD's off the SATA3 for the SSD(s) only and the HDD's on the SATA2. The SSD isn't like a HDD and 1/100th filled or 99/100th filled makes no difference.

Watch this example {opening Apps} http://www.youtube.com/watch?v=6jsHpNkDQn0

 

[Proximon]

Hibernation is not sleep. Hibernation is were the current system info is written to the drive, so that the entire system including RAM can be powered down. It's only used in laptops and really has no use in a desktop at all. It's just a very few watts to keep the RAM up and holding the info.

I don't think system restore is a big thing, but I do believe what Doug said: "And for all of the value that gets heaped on System Restore, it's not always able to facilitate a clean recovery."
It really isn't. I have had it fail more often than not. I just worked on a laptop that I had previously worked on, where I made sure System Restore was running. The OS was, once again, corrupted (porn and idiocy are a bad mix) and I tried to roll back the system. There were 6 restore points and NONE of them worked, but a format and re-install worked fine.

So these days I prefer a full back-up. Plus I have space on my SSD for stuff I want there, like Civ V. It's amazing what will fit on a 160GB drive if you keep it cleaned up. 9 steam games, and none of them small. Windows Virtual PC, a local MySQL server (just for learning).

Granted, I just have 25GB free on the SSD atm, but there are downloads I can remove.

 

[jaquith]

Say in another life I understood the hacker - trojans/worms, etc and exploits on servers. The reason that "There were 6 restore points and NONE of them worked" failed is the first thing a hacker does is to corrupt {re-write} the registry and corrupt 'System Restore'; most hackers change '.exe' entries to re-install and re-run their 'Appz'.

I've had NIS & NOD32 miss Java exploits of all places Java banner ads on SuperPages.com - the first thing I do is Ctrl+Alt+Del and look for the Appz, kill it, and immediately fix {.exe} entries, and also periodically save a copy of the registry. I wrote something to compare before & after to see the 'changes.'

To be clear I never said "Hibernation is sleep", but Hybrid Sleep is a cross between Standby & Hibernation {S3 + S4}.

The {hiberfil.sys} is used for both Sleep Mode and Hibernate mode; so if you delete the hiberfil.sys you lose both Hybrid Sleep {S3+S4} and the ability for the PC to Hibernate. Conversely, if you want either Hybrid Sleep <or> Hibernate you need the hiberfil.sys file. {hiberfil.sys} is the saved state of Memory.

 

[flong]

Say in another life I understood the hacker - trojans/worms, etc and exploits on servers. The reason that "There were 6 restore points and NONE of them worked" failed is the first thing a hacker does is to corrupt {re-write} the registry and corrupt 'System Restore'; most hackers change '.exe' entries to re-install and re-run their 'Appz'.

I've had NIS & NOD32 miss Java exploits of all places Java banner ads on SuperPages.com - the first thing I do is Ctrl+Alt+Del and look for the Appz, kill it, and immediately fix {.exe} entries, and also periodically save a copy of the registry. I wrote something to compare before & after to see the 'changes.'

To be clear I never said "Hibernation is sleep", but Hybrid Sleep is a cross between Standby & Hibernation {S3 + S4}.

The {hiberfil.sys} is used for both Sleep Mode and Hibernate mode; so if you delete the hiberfil.sys you lose both Hybrid Sleep {S3+S4} and the ability for the PC to Hibernate. Conversely, if you want either Hybrid Sleep <or> Hibernate you need the hiberfil.sys file. {hiberfil.sys} is the saved state of Memory.

OK guys, since this subject has been broached and I have two computer masterminds here 🙂, what is the best sleep mode for my desktop? Currently I am not sure what sleep mode it is in when it sleeps but I did try to set it for hybrid mode because I had read that was the best.

Also, since I have the best of the best here, is there a sleep mode that will put the computer to sleep while at the same time allowing Windows Media Center to record a program from my HDTV card? Right now it won't go to sleep when recording and it keeps the computer fully on. I would like to find a sleep mode that allows WMC to record the program while the computer is asleep - if that is even possible.

 

[jaquith]

TIP: Open regedit and KNOW these primary keys:
1. HKEY_CURRENT_USER\Software\Classes\.exe\shell\open\command
(Default) REG_SZ exefile
Content Type REG_SZ application/x-msdownload

2. HKEY_CLASSES_ROOT\.exe\shell\open\command
(Default) REG_SZ exefile
Content Type REG_SZ application/x-msdownload

3. HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Security Center "AntiVirusOverride" = "0"
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Security Center "FirewallOverride" = "0"

See - http://www.symantec.com/connect/blogs/xp-internet-security-2010-rogue

BTW - Chances are your System Restore 'was' good, but the Appz once you ran regedit corrupted the restore file(s). 😉

 

[caesparktom]

Best answer selected by caesparktom.

 

[jaquith]

I'd guess none - no sleep mode, if you're recording then the PC must be ON {S0}, but other times use the Hybrid Sleep which is Window's Default sleep.

 

[caesparktom]

Wow you are the first person I have heard that has said system restore should be enabled. I will do some more research because I personally like the MS System Restore function A LOT. It is a great feature in Windows 7 and it works very well. It has saved me a couple of times. Hibernation is not very useful to me because my computer does go into hybrid sleep mode but it actually takes just about as long to put it to sleep and start it up as it does to turn it off and start it up. It boots Windows 7 in less than 25 seconds (from the point that you push the ON button).

Right now I have all data going to a 1 TB Samsung F3 and another Hitachi drive. I did not move the Outlook PST files to the F3 (they are on the SSD) but that probably is a good idea considering the size of the email files. I think I can move it without having to reinstall Outlook but I will have to look into that.

BTW what are OST files? Operating system files? Should I move them off of my SSD?

I use the Force 3 GT 120GB drive only for program files and the operating system and I don't even have it half filled. With this particular drive the performance degradation between 1/2 - 3/4 fill is insignificant and so I do not have to worry about the 1/2 fill point. My goal is to keep it near the half fill point so that I do not lose performance.

That being said, my Samsung and Hitachi ATTO scores are very close to 150 mb/s read and write speeds. So they are no slow pokes by any means. I was blown away that both of these HDDs pushed the 150 mb/s mark - the Hitachi is a 6 gb/s drive and it is connected to a SATA III port (I have extra SATA III ports). The Samsung is SATA II but it is almost as fast as the Hitachi.

yeah, that restore point one was really a shocking/irritating. Also very hard to skip when they begin: "This is the most important piece of information in this Guide...Typically, System Restore will degrade SSD performance significantly within a few weeks and makes it very difficult for TRIM to function. Disabling System Restore is not only recommended throughout the web, but also, Intel has lately also included this advise with respect to its ssds." - Yikes!

flong are saying you only plan to use half the drive capacity? What kind of performance degradation can i expect past "1/2 fill point"? btw, how do you monitor you read/write speeds?

 

[flong]

yeah, that restore point one was really a shocking/irritating. Also very hard to skip when they begin: "This is the most important piece of information in this Guide...Typically, System Restore will degrade SSD performance significantly within a few weeks and makes it very difficult for TRIM to function. Disabling System Restore is not only recommended throughout the web, but also, Intel has lately also included this advise with respect to its ssds." - Yikes!

flong are saying you only plan to use half the drive capacity? What kind of performance degradation can i expect past "1/2 fill point"? btw, how do you monitor you read/write speeds?

Hi 🙂, I plan to let the drive fill to no more than 3/4 capacity. I would like to keep it closer to 1/2 fill capacity. Right now I have pretty much everything I want on the drive and I am not even near 1/2 capacity.

This is important because the drive performance is affected by how much is on the drive. Different reviewers test this performance - often it is call "fill performance" or something close to that. The problem with SSDs is that depending on the flash memory used, and the size of the NAND modules, the performance of the drives drop precipitously with the amount of data stored.

For example, the OCZ 3 240 GB Agility's performance falls like a rock when it is filled because it uses a cheaper flash memory. Once it is close to half full, its performance drops to something roughly equal to the second generation Sandforce drives (or less) even though it is a third generation drive. This makes a 240 GB drive much slower than your V3 or my Force 3 GT. This should not occur because you pay a premium for the 240 GB Agility. When fill performance is considered the 240 GB Agility is a failure when cost vs performance is considered. However, take heart, the OCZ V3s are in another world of quality and performance - they are much better than the Agility series.

ALL SSDs degrade in performance as they are filled, even the best of the best. The question then becomes what is the sweet spot of fill capacity vs performance. Fortunately, your V3 has a very good "fill" performance. I think (and this is from memory) that you can fill your drive pretty much the same as mine up to 3/4 full and still maintain most of your performance. I plan on keeping mine around half-full simply because that is all that I need and I have a 1 TB and 1.5 TB HDD for other things that I want to store. For example, all of my WMC files and recorded HDTV programs are stored on the 1 TB Samsung drive.

To monitor my SSD speed I downloaded ATTO software - it is free and it is heavily used in reviews. It is also the test that manufacturers most often use for their SSD read and write claims. Just Google ATTO and you will find it.

For now I have disabled System Restore on my SSD. I am checking into Mozy or another cloud backup service since that is the most secure way to back up data anyway. I am not sure if those services can restore the operating system - I have heard that the cannot.

Still if I had to do a clean install on my SSD, it would not be overwhelming because all that I have on it is the operating system and programs. Yes, it would take longer than System Restore by quite a bit, but System Restore does not always work. I am researching Jaquith's data to try to understand the issue better. For now I cannot allow System Restore to degrade my SSD's performance or life span - this could probably not be resolved with a clean install.

Jaquith (and Proximon) is far more experienced with computers than I am right now and he understands a level of the System Restore that I do not. His point that Trojan worms and spyware must be disabled before the restore is quite brilliant - but I am not sure that I would be able to recognize which EXE program that is running that is the spyware that is causing the problem. A clean install DEFINITELY would solve the problem. Really, if I had a severe spyware or virus problem I would do a clean install anyway just to be safe. It is kind of like making sure all of the fleas are killed on your dog before letting him back in the house ha, ha.

I am very careful with my downloading and surfing and I have never had spyware or a virus take my computer down (I have been attacked but Avast is very good at shutting down attacks and I run MalwareBytes anti-spyware regularly as just a check (it is freeware and it works). OTHER problems with Windows can cause the computer to fail - most often Windows 7 files become corrupted for various mysterious reasons. For those problems, the System Restore is quite handy, but it does not always work.

This is my first SSD and so I am still in the learning curve. I am VERY satisfied with the Corsair Force 3 GT 120 GB SSD, it is putting out performance that competes with many 240 GB drives. The V3 is also a top performer.

 

[jaquith]

IMO BS, here's the proof. Both these tests are from my wife's PC {same to same} but over a full year+ later. Further, on her's nothing is or has been 'tweaked', it's 100% Windows Default and a huge 30GB OST. Oddly, if anything it's faster.

8/22/2011

5/20/2010

 

[flong]

I forgot to mention that right now even though my SSD is approaching half full (I have 77.7 GB free of 111 GB) my performance has not diminished according to ATTO. I was something like 560 - 568 mb/s read and approximately520 mb/s write when the drive was brand new. Right now I am at 557 mb/s and 515 mb/s respectively. Not a significantly measurable difference really.

 

[flong]

IMO BS, here's the proof. Both these tests are from my wife's PC {same to same} but over a full year+ later. Further, on her's nothing is or has been 'tweaked', it's 100% Windows Default and a huge 30GB OST. Oddly, if anything it's faster.

8/22/2011
http://i1013.photobucket.com/albums/af254/Jaquith/ATTO_HJ_SSD.jpg

5/20/2010
http://i1013.photobucket.com/albums/af254/Jaquith/RAID_SSD_ATTO.jpg

That is fascinating - have you taken into account any firmware updates? I am assuming that these are second generation SSDs by their speeds? Also these are probably 34 nm NAND right? They would be affected less by re-writes if that is the case. Also, for someone who does not use their computer a lot, it would take longer for fill degradation to appear.

Don't get me wrong, I am not saying I am an expert with SSDs. I am just stating what I have found as I have researched the issue. Anandtech is the principal reviewer that talks about "fill" performance and they are pretty well respected. That is why I am checking the ATTO benchmarks on my SSD as it fills. I would like to get some first-hand experience with the issue. So far I have not experienced ANY significant performance degradation as my SSD has filled. I just checked mine again and it was 555 mb/s read adn 510 write. But that is comparing it to its brand new state - about 10 mb/s drop; not much.

Thank you for the first-hand experience - it is invaluable. I think there is a big learning curve right now with SSDs - even among the experts.

 

[jaquith]

You need to separate Myths from Facts and you'll be ever more happier. In that article that Doug wrote I had already tested the supposed 'tweaks' and discovered them to be Fantasy Land. Often SSD improve after a few R/W cycles. There are so many other background constant R/W that it would make your head spin-off.

Regarding TRIM Myths, it works in either AHCI <or> IDE and about the only TRIM 'killer' is RAID. The lifespan of 'YOUR' SSD is the last worry you need, all of my Primary SQL Servers are running RAID with SSD and I don't even give it a thought. My use in a month surpasses what you'll ever do on your PC.

Noted, some early SSD's were crap. You all need to read this article and quit worrying about frigging write-cycles another Myth in real terms/world -> http://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923.html If you run out of space - easy fix: 1. don't put important data on it and RAID 0 (2) SSD; twice the size and nearly twice as fast, SSD's scale pretty good.

Again, quit babying your SSD - just use it.

 

[flong]

You need to separate Myths from Facts and you'll be ever more happier. In that article that Doug wrote I had already tested the supposed 'tweaks' and discovered them to be Fantasy Land. Often SSD improve after a few R/W cycles. There are so many other background constant R/W that it would make your head spin-off.

Regarding TRIM Myths, it works in either AHCI <or> IDE and about the only TRIM 'killer' is RAID. The lifespan of 'YOUR' SSD is the last worry you need, all of my Primary SQL Servers are running RAID with SSD and I don't even give it a thought. My use in a month surpasses what you'll ever do on your PC.

Noted, some early SSD's were crap. You all need to read this article and quit worrying about frigging write-cycles another Myth in real terms/world -> http://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923.html If you run out of space - easy fix: 1. don't put important data on it and RAID 0 (2) SSD; twice the size and nearly twice as fast, SSD's scale pretty good.

Again, quit babying your SSD - just use it.

Thanks again for your advice. As an experienced user of SSDs it is invaluable. Thanks for the heads up also that SSDs can be used in RAID 0 and scale well. That is exciting.

Keep in mind that the OP and myself are new to SSDs and so we are in the process of trying to separate "myth from fact." That is always the challenge with any complex subject. I have read nearly every expert review avaliable, but even the experts are disagreeing now. Since you have high-use first-hand experience, I really value hearing your opinions. Are you using enterprise grade or regular user SSDs like me and the OP? That would also make a difference.

I definitely agree with you that that many of the SSD tweaks listed in that article are unnecessary and unproven - the article does a pretty good job of qualifying that.

 

[flong]

Jaquith, following is the "fill" testing in the Tweaktown review of the 240 GB Agility 3. You will note a significant drop in performance when compared to other third generation Sandforce drives. This is what I was talking about.

http://www.tweaktown.com/reviews/4139/ocz_technology_agility_3_240gb_ssd_review/index10.html

In most cases, the Agility's 240 GB is performance drops almost 40% more than the Vertex 3 120 GB performance as the drives fill. ALL of the drives lose performance as they fill.

 

[jaquith]

From your prior question, no Firmware upgrades. I just knew I had an old ATTO screenshot handy and I wanted to show you a Then vs Now comparison. I was so convinced they'd be the same that screen shot & bench was posted within 5 minutes of running ATTO.

Yes, I am using Intel Enterprise SSD of various capacities. Been looking at OCZ...verdicts still out. I own an REO/IDX/TAX data center; think about how many parcels and subsequent data there is in just (1) one state, now relate all those different records and tables and you'll begin to understand the shear amount of raw data.

As I said, people parrot what they read to be the truth. I on the other hand question everything and need to 'see' it for myself. There are sooo many Skewed/Biased benchmarks it's best to see overlapping results. I tend to argue because I also hobby build so I see the new stuff all of the time. I have (3) Articles that I'm thinking about right now that I simply shake my head.

There's a difference between a 'Paid' Expert and 'Disinterested' Expert. Paid Reviews and User Reviews - if I know the guy and know he/she is competent I'll believe the User Review. While I'm thinking about it, be leery about Newegg 'Feedback' filter the 'verified' and see if any parroting is going on and skip those reviews. Further, some of the best stuff has no Feedback - it's either new or expensive.

IMO - long-term just let Windows do it's thing, Windows 7 unlike Widoze Vista is pretty good, but nothing is perfect.

 

[jaquith]

This is what I was talking about.

http://www.tweaktown.com/reviews/4139/ocz_technology_agility_3_240gb_ssd_review/index10.html

In most cases, the Agility's 240 GB is performance drops almost 40% more than the Vertex 3 120 GB performance as the drives fill. ALL of the drives lose performance as they fill.

I will look at this in the morning, it does no good unless I read it, confirm it but I'm tired. I'll reply. If it was written by Lsdmeasap then you can believe it.

I just know if I had a 10% much less a 40% drop-off I'd consider the drive defective. Ask yourself this question, do you get a 10%~40% drop-off with your RAM? On the servers, I have OS and RAID SSD drives. OCZ use to make the worst RAM you could buy...reason they quit producing RAM. Looking at the Corsair no difference...ditto with my Intel. This could easily be explained by a firmware issue for the controller.

 

[flong]

I will look at this in the morning, it does no good unless I read it, confirm it but I'm tired. I'll reply. If it was written by Lsdmeasap then you can believe it.

I just know if I had a 10% much less a 40% drop-off I'd consider the drive defective. Ask yourself this question, do you get a 10%~40% drop-off with your RAM? On the servers, I have OS and RAID SSD drives. OCZ use to make the worst RAM you could buy...reason they quit producing RAM. Looking at the Corsair no difference...ditto with my Intel. This could easily be explained by a firmware issue for the controller.

Note that Tweaktown does fill testing on every SSD it reviews and they all lose performance as they fill according to their benchmarks.

Thanks, I am looking forward to your opinion on the article. I value this reviewer (TweakTown) because they try very hard to mimic the worst-case real world use. Kind of like putting a computer through Prime 95 testing to see how it holds up.

They also use real-world benchmarks with common applications for their fill testing instead of synthetic benchmarks. It makes their benchmarking more credible to me.

For the Agility, the Tweaktown review actually shows over a 50% drop in performance when it is 3/4 full. Note that not every reviewer looks at the fill performance of the SSDs that it tests.

 

[caesparktom]

have a mundane questions regarding file managemnt. Is keeping os and aps on ssd and documents on hdd primarily an issue of storage. I had intended to store docs on hdd but use sdd for working files. Is that good practice and is there any reason to relocate My Documents/User folder off ssd then? Also, i typicaly manage a new folder for documents rather than moving My Documents because I don't like how applications put stuff in that folder without asking me. So basically I create a folder called C:\mydocs (easier to type and not buried) and leave My Documents where it is and let programs do what they like with it.

This is what I was talking about.

http://www.tweaktown.com/reviews/4 [...] dex10.html

In most cases, the Agility's 240 GB is performance drops almost 40% more than the Vertex 3 120 GB performance as the drives fill. ALL of the drives lose performance as they fill.

this is truly startling if the study is good and these aren't defective drives that were tested. perhaps i should keep the drive completely empty-- like a sweet car that doesn't leave the garage Seriously though. I am going to use as much as i need in the spirit of jaquith's "quit babying your SSD - just use it" advice and look forward to testing this phenomena myself.

btw, i have reluctantly re-enabled system restore.

 

[flong]

have a mundane questions regarding file managemnt. Is keeping os and aps on ssd and documents on hdd primarily an issue of storage. I had intended to store docs on hdd but use sdd for working files. Is that good practice and is there any reason to relocate My Documents/User folder off ssd then? Also, i typicaly manage a new folder for documents rather than moving My Documents because I don't like how applications put stuff in that folder without asking me. So basically I create a folder called C:\mydocs (easier to type and not buried) and leave My Documents where it is and let programs do what they like with it.

This is what I was talking about.

http://www.tweaktown.com/reviews/4 [...] dex10.html

In most cases, the Agility's 240 GB is performance drops almost 40% more than the Vertex 3 120 GB performance as the drives fill. ALL of the drives lose performance as they fill.

this is truly startling if the study is good and these aren't defective drives that were tested. perhaps i should keep the drive completely empty-- like a sweet car that doesn't leave the garage Seriously though. I am going to use as much as i need in the spirit of jaquith's "quit babying your SSD - just use it" advice and look forward to testing this phenomena myself.

btw, i have reluctantly re-enabled system restore.

I do have first hand experience with data storage on my HDDs. With my programs stored on the SSD and the data from the programs stored on HDD, I get instantaneous load times. Now you if you have a real data heavy files (hundreds of MBs) then it might be different. Storing data on your HDD should not slow you down at all and it will save room on your SSD for more important things. It is easily switched if back to the SSD if needed.

I believe that the Tweaktown reviews are accurate. Only 2 or 3 sites that I can think of do fill testing. Note that Tom's Hardware also does a kind of "fill" testing analysis and they acknowledge the performance degradation. I am somewhat surprised that Jaquith says that filling the SSD has no effect. However, since I am a novice, I am willing to keep an open mind, humble and learn from more experienced users.

So far with my SSD over 1/3 full I can measure very little performance degradation (10 mb/s) but my drive is brand new.

 

[caesparktom]

hmm. well 100+MB very common. 500+MB not at all uncommon. I will do a test and let you know if/how 500+MB load time differ.

 

[jaquith]

Okay, some quick looking. This issue happens IF you rig the test, and by rigging I mean filling the drives with incompressible data. In particular the issue is more from SandForce SF-1200 than the SandForce SF-2500 by the SSD being bottlenecked by its block recycling path. Furter, you need to also give TRIM time to 'catch-up' rather than testing in some oddball state.

So, in the 'REAL WORLD' you'll never see this problem -- As I said be leery of articles and the fine print.

 

[flong]

hmm. well 100+MB very common. 500+MB not at all uncommon. I will do a test and let you know if/how 500+MB load time differ.

I will not be handling the size of files that you will but it is easy for you to test. I would suggest that you store a 500 mb file on your Samsung and then open it with the accompanying program stored on your SSD. I am betting that it will open very fast. If it doesn't you can easily switch your file storage over to your SSD.

Think about this also - your file management will be much easier if you organize and store your data files all on your HDD. Trying to separate some files to go to the SSD and some to go to the SSD will in itself cost you valuable time.

Let me know what you find out, I am curious. I am betting that a 500 mb file stored on your HDD will open very quickly. So far with my system I cannot find any file stored on my HDD that does not open instantly but they are not 500 mb in size.

 

[flong]

Okay, some quick looking. This issue happens IF you rig the test, and by rigging I mean filling the drives with incompressible data. In particular the issue is more from SandForce SF-1200 than the SandForce SF-2500 by the SSD being bottlenecked by its block recycling path. Furter, you need to also give TRIM time to 'catch-up' rather than testing in some oddball state.

So, in the 'REAL WORLD' you'll never see this problem -- As I said be leery of articles and the fine print.

Thanks for looking at the review. It seems that in the world of SSDs even expert reviews may not be reliable. The Tweaktown review definitely states that the performance of SSDs will degrade as they are filled. I am always leery of the information that I read and I try to verify it from different sources. As I mentioned above, there are at least 3 sites that do a type of "fill" testing and all of them report performance degradation - including Tom's Hardware. I don't say this to infer that you are wrong. I am saying this because it is evidently part of the "myth" that I must sort through.

I will be interested to see how my SSD performs over time. So far my first-hand experience confirms what you are saying. Based on the fill test in Tweaktown, my drive should have lost at least 15% - 20% of its performance. So far that has not been the case.

One of the biggest problems I have found with professional reviews of computer components is that they often do not mimic real-world situations (if there is a such a thing with such a base of diverse users). Taking the information in the reviews and distilling the information to see how it applies to a particular situation (my own build for example) can be very challenging.

For example, when I was researching my case, the 650D, two professional review sites said that it was "loud." Every other professional review said it was "very" quiet. Go figure. Now that I have the case, with the fans on high you can hear them, but it is because they are moving air. The case is quiet on full fan; it is about the same as my I-7 920. With the fans on low, the case is nearly silent.

The challenge is sorting through the reviews to figure out the what they actually mean. That is why I am not surprised to hear that SSDs are more durable than they have been portrayed.

 

[jaquith]

...The Tweaktown review definitely states that the performance of SSDs will degrade as they are filled. I am always leery of the information that I read and I try to verify it from different sources. As I mentioned above, there are at least 3 sites that do a type of "fill" testing and all of them report performance degradation - including Tom's Hardware...

Yes, the $100 question is filled with what! If you fill it with tons of H.264 files or other incompressible data then yes, most everything else no.

I could go into all sorts of 'CrAzY' arguments I've had with both Reviewers and people posting all sorts of stuff.

If I drop a 30# hammer on your case and it breaks does it make it a bad case, or does it prove a moronic test.