monster707
Member
4 posts | I found a link on another forum, for the 3AAE firmware, need to back up my files before I try it. don't know if its against the rules to post links. Credit goes to original poster "Presjar" in the other forum. From the name of the file, I assume it's only for the seagate 320 gig drive
http://depositfiles.com/en/files/1626013
just remove the xx's and add tt
monster707
//Admin edit: you can safely post links here  |
Omion
Unregistered | Thanks for the link, monster707!
Unfortunately, though, the firmware update doesn't work. It claimed to flash, but it still says 3.AAK and has the same performance issue. Oh well. |
Madman
Unregistered | I have 3 AAJ and 1 AAE ST3500630AS drives. They were purchased a while ago, the AAE first then the 3 AAJ all together a little bit later. The AAJ drives exhibit the same exact HDTune and to a lesser extent HDTach performance drop compared to the AAE but I've only realized this with recent testing. I had originally tried to get all AAJ from SEagate assuming that newer firmware was better  but after two tries of getting back AAE they finally told me that 'We cannot guarantee replacement with certain firmware' apparently two reps in a row weren't aware of this policy *cough* while they were nice I think they just wanted to keep their call times fast.
My question is despite the synthetic results is it possible this isn't as big an issue as we're making it?
Enlightenment wrote: The difference between HDTune and ATTO is, that ATTO tests on the filesystem (NTFS/FAT) and thus enjoys optimizations such as read-ahead and write-buffering. HDTune on the other side, is a strongly synthetic benchmark that reads directly from the device. The scores of HDTune are not representative of real performance, although in this case it did produce an abnormal figure which leaded to suspicions and finally my article.
The scores from HDTune are low because it reads with a blocksize of 32KB. When looking at the DD-tests we can clearly see the AAK performing poorly with such low blocksizes. For real performance tests, you need a more advanced benchmark than either HDTune or ATTO.
By 'more advanced test' do you mean real-world app tests? Because in many of those the 3AAK does do well aside from the last one. |
_Shorty
Member
6 posts | I phoned Seagate's tech support number to ask about this. As soon as I started talking about firmware the 1st-level tech support guy escalates me without asking anything else. The 2nd-level guy does a bit of reading and seems to think this AAK firmware is an OEM firmware, and that Seagate isn't obliged to do anything for me at all. I'm told to contact the store I bought it from, as it is an OEM drive and the OEM is responsible for any support or replacement options, etc, etc. What a joke. He says the AAE firmware is the latest firmware for Seagate's retail drives, and was a bit confused at first that this drive had AAK firmware. He ends up talking to one of the head engineers real quick, and when he comes back he relays that apparently some OEM company dumped a whole wack of drives onto retailers. Seagate basically told me good luck, but we ain't gonna do squat for ya, sonny. Lovely. |
neutronium
Unregistered | N00bit question on these drives.
Mine (AAK versions) still have the jumper on, but the machine's SATA controller appears to be an early one, (only running in UDMA-5 mode., the drive supports UDMA-6).
I can't imagine any benefit from pulling the jumper, correct? |
Enlightenment
Administrator
 105 posts | @neutronium: the jumper controls the SATA 1.5Gbps / SATA 3.0Gbps transfer mode. If you remove the jumper, you will run in 3.0Gbps mode, sometimes wrongly referred to as SATA2.
You can only benefit from this is your controller also supports 3Gbps operation, and even then the performance benefit is quite small. If your controller is an older one, just keep the jumper on. Else it is possible your controller will not detect your hard drive. Take control of the input and you shall become master of the output. |
neutronium
Unregistered | Thx Enlightenment. I had reasoned it to there as well and didn't feel like shuffling things around unless I had to |
Kboxvegas
Unregistered | I have two Seagate 500GB 7200.10's with the AAK firmware bought 2 months apart from one another. One I got OEM from newegg and the other from fry's in the older "big" retail box and both were made in thailand. I was curious after reading about this whole firmware issue so I ran some test with a biostar tforce965PT board using the ICH8 controller in sataII "jumpers removed from drives".
I dont know why but I am not seeing this "capped" effect.
 
  |
monster707
Member
4 posts | @ Kboxvegas
The reason your not seeing the cap issue, could be that your using the new hdtune 2.54, which now has the option to change the block size of the test, default is 64kb, try using the old hdtune verion 2.53, and then you might see the "capped effect". |
Snake
Unregistered | I've got three AAK drives as well 
The shop i bought it from was so nice to exchange them for AAE drives, so now i'm happy again.
And indeed my RAID array scores as much faster now (i've got 4 seagate disks in RAID0). On the Intel ICH8R controller.
Thanks for your article, i appreciate it! |
DiGiTaL_MoNkEY
Unregistered | For people testing their drives now the new HD Tune (2.54) it uses a DEFAULT block size of 64k (you can change the size in the new version) as compared to the previous version 2.53 which used a 16k block size which could not be changed.
From some peoples "new" tests found that when on the 64K block size the drive runs as expected, when using the 16k block you have the low preformance effect.
It may of been done since files are getting bigger & this firmware works best or better with bigger block sizes, so maybe when formating a drive if using NTFS use a 64k or a 32k "allocation unit size".
Will be getting some new 7200.10 hard drives this week & will post the results.
|
Paulo
Member
7 posts | Hi all.
I have an 500GB AAK drive and *now* I really don't feel too bad about it, just because it was inexpensive ( 99€ ) and has 5 years warranty.
And then maybe I prefer it to the AAE.
I took a look at your article. Let me say that it is fantastic and insightful, it's not a "me too" review.
Since I was annoyed with my drive, I read the article carefully. It was interesting that the AAK drive is faster on the so called "real world" benchmarks.
Unfortunately the PCMark2005 graph don't show the real difference between the drives because of the range of values, the large bars compress the smaller too much
By percentage (AAK vs AAE):
Virus Scan: 1,3%
File Write: -9%
XP Startup: 14.8%
App Loading: 9.2%
General Use: 11.9%
IoMeter: 4~5%
h2benchw: basically the same, didn't bother tbh.
WinRar: -12.5%
Recopy: -11.5%
See my point?
That leaves the sequential file read.
Honestly I didn't test both on Vista and XP, but on Vista, file copies use 1MB block transfers
http://en.wikipedia.org/wiki/Transactional_NTFS#I.2FO_subsystem
Prior to Windows Vista, all I/O requests were capped at 64 KB; thus larger operations had to be completed in chunks. In Windows Vista, there is no limit on the size of I/O requests. This means an entire I/O operation can be completed by issuing fewer requests, which in turn means higher performance. Windows Explorer copy engine issues 1 MB requests in Windows Vista.
Theoretically this drive performs much better on Vista than XP. Or is it the other way around, cripples XP on purpose? Does it quack like a duck?
I mean, what's the technical reason for this piss poor performance at slower block sizes? I have an Hitachi 250GB VLA and a Samsung 1614N P-ATA and they perform much better at 4k (using dd)
Now, I see that this drive is adequate for Vista bandwagon and Office PCs. Also with lower OEM margins, Seagate cut corners and AAK drives may have slower electronics.
So, in the end, I really dunno which one is best for me, using XP and being also a gamer. XP loads faster, some general work may be faster, but map loading can be way slower.
Otoh, if you use Vista and don't play games or edit video, don't feel like you bought a dude and don't bother. The drive may be more adequate than the AAE better for you
|
DiGiTaL_MoNkEY
Unregistered | Had a little chat with Seagate Live Support:
May have to give tech support a physical call & see what they have to say.
__________________________________ ___________________
You: and i am wondering about the details of the AAK firmware on this recent drives and it may effect the performance of the actual hard drive
You: I've seen quite a few tests that it does not come close to the 78mb/s transfer rate, most closer to the 60mb/s mark or even less.
Kamel TSXXX: AAK firmware will only show the performance decline when used by Bench Marking software and not in real life applications
You: Thanks for answering my question, is there any chance that any information is avaliable on the firmware changes in this release as compared to maybe the AAC version??
Kamel TSXXX: No, as it does not go backward
Kamel TSXXX: But
Kamel TSXXX: The 7200-11 are being pushed the market now
|
cjc1103
Unregistered | I just got off the phone with a Level 2 tech at Seagate. I have 8 of these 500GB ST3500630AS (7200.10) drives, all with the 3.AAK firmware. The AAK firmware is on the newest controller boards manufactured in China, and there is no way to change/flash the firmware. He said they could not even do it in house, the board does not have flash capability. He downplayed the performance difference between AAK and the previous firmware revisions, but I saw in the post above the AAK firmware is apparently optimized for 64k block size. The only thing I can do is RMA the drives, and I would get back an identical drive, which may or may not have the AAK firmware. Since I am running these drives in a RAID5 setup (2 servers with 4 drives each), I want the same firmware for all the drives, so I'm going to stick with what I have. I don't know whether I would even see a performance difference with an older firmware version/drive in a file server, the network is usually the bottleneck even with GbE, and these servers are lightly loaded anyway. But I am going to change the block size on the filesystem to 64k, it can't hurt, since most of my files are huge (>100MB). Also I am going to remove the jumper to force SATA II mode, since I have a SATA II interface in each server.
- Chris |
DiGiTaL_MoNkEY
Unregistered | Had ANOTHER chat with the live customer support person and i had to drag the questions out of him...some of of the lines have been removed, just the important stuff was left..
I was i right about the algorithms being different (as i thought to myself). Also i never asked him about HDtune he was the one to mention it
__________________________________ _________
Rob: And i'm been reading quite a bit about the performance i would get with these drives.
Rob: I have found that MANY people having issues with the latest AAK firmware, should i be concerned about this?
TechSupport: No
Rob: They are stating that the performance has been reduced as compared to the AAE firmware drives, is there any details about the firmware changes in the AAK drive?
TechSupport: No it has increased. The tool that they use HDtune can not read our new firmware correctly. This should be resolved in the next revision of HDtune.
Rob: Is there any benchmark software that you would recommend to test these drives??
TechSupport: We do not recommend using any software, Use real work things like transferring files to the drive and then doing the math.
Rob: Is there any possibility to elaborate why this benchmark does not read the 7200.10 drives correctly?
TechSupport: The algorithms in the firmware are not written to satisfy synthetic benchmarking programs. So the only way to see the true performance is to do real word test.
__________________________________ ______ |
jealma
Unregistered | For a RAID-setup I'm at the point of buying 8 7200.10 750GB drives. These drivers will be hooked up to an Areca ARC1220. Since it seems to me that the drives with AAK firmware have better real-world performance and the AAE firmware does better in terms of sequential reads and writes; what is the smart choice to make? I can make a request for drives with an AAE firmware, or with AAK firmware, but what will be better for me? These drives will be in a RAID5 array (1 hotspare, 7 in RAID5) and it will mainly contain movies and series (files of around 700 MB). |
Enlightenment
Administrator
105 posts | DiGiTaL_MoNkEY wrote: TechSupport: The algorithms in the firmware are not written to satisfy synthetic benchmarking programs. So the only way to see the true performance is to do real word test. And that's what i did, using explorer (sequential) and WinRAR (non-sequential). All real-world tests showed that AAK is quite a bit slower than AAE.
jealma wrote: Since it seems to me that the drives with AAK firmware have better real-world performance I'm not convinced about that. While it's true the application benchmarks all favored the AAK drive, i can't tell how they work or re-produce their results in any real-world test. I did tests with extracting many small files (108 thousand to be exact) but all these tests favored the AAE drive over the AAK.
It is true though, that HDTune is a pure synthetic application and by itself does not say anything meaningful about real-life performance. Take control of the input and you shall become master of the output. |
shopaholic
Unregistered | I just got a 500GB 7200.10 AAK that was made in China. Can't wait to hear updates. A lot of people claim the ones from China are loud too. Guess I'll wait and see. |
hudson
Unregistered | The Seagate 320 gig drives from China are getting worse. I bought 4 from two different suppliers in 2 different cities - they all failed within 24 hours. Seagate doesn't want to admit there's a problem with the old China Maxtor factory, and will only replace the drives with used remanufactured drives of the same make and model.
They won't even give me the address to send the court papers to sue - not by phone, not by email. Cowards.
More here: http://slashdot.org/~tomhudson/journal/184071
The first retailer says that Seagate won't take any returns from them any more - all returns must go directly to Seagate (in other words, they had too many returns - gee, I wonder why?) The second only wants to swap for "more of the same."
As far as I can tell, the 5 year warranty is the same scam muffler shops pull - they figure most of the replacement drives will go into machines that will be obsoleted before the 5 years are up. |
sasa
Unregistered | Sander wrote: Hey! Great article!
Ofcourse i'm also very interested in the response from Seagate. I got two AAK drives and both exhibit the same problem as described in your article.
I also got loud beeps every hour or so, did you notice those too? |
