The USB-Stick Benchmark Thread
#26
Posted 22 June 2010 - 03:54 PM
Those sticks were mostly just bigger versions of sticks, everyone already had and liked, from the handling, the features, the performance.
I had to send back every single one, except my SanDisks. Because of the unbelievable poor performance compared to their older counterparts. Not just theoretical, but noticeable during use.
The best one delivered just ~60% of the performance of the older one, the worst only ~40%!
Talking to the support of the later one, i got as a reply, that their drives are sold unrated and tolerances would be normal.
Does anyone really consider 60% off, still a tolerance?
Anyway, watch out! Lot's of crap sold right now, it seems.
#27
Posted 22 June 2010 - 04:24 PM
capacity 8069840896 bytes
-----------------------------------------------------------------------
CrystalDiskMark 3.0 x64 © 2007-2010 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
Sequential Read : 20.827 MB/s
Sequential Write : 5.717 MB/s
Random Read 512KB : 20.812 MB/s
Random Write 512KB : 1.604 MB/s
Random Read 4KB (QD=1) : 6.324 MB/s [ 1543.8 IOPS]
Random Write 4KB (QD=1) : 0.020 MB/s [ 4.9 IOPS]
Random Read 4KB (QD=32) : 6.083 MB/s [ 1485.1 IOPS]
Random Write 4KB (QD=32) : 0.028 MB/s [ 6.8 IOPS]
Test : 50 MB [H: 0.0% (0.0/7676.9 MB)] (x1)
Date : 2010/06/22 23:17:36
OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x64)
#28
Posted 22 June 2010 - 05:09 PM
Ordered a bunch of USB-Stcks for my family.
Those sticks were mostly just bigger versions of sticks, everyone already had and liked, from the handling, the features, the performance.
I had to send back every single one, except my SanDisks. Because of the unbelievable poor performance compared to their older counterparts. Not just theoretical, but noticeable during use.
The best one delivered just ~60% of the performance of the older one, the worst only ~40%!
Talking to the support of the later one, i got as a reply, that their drives are sold unrated and tolerances would be normal.
Does anyone really consider 60% off, still a tolerance?
Anyway, watch out! Lot's of crap sold right now, it seems.
I had a similar issue with two 2GB Lexar Firefly Jumpdrives. Supposedly identical models (they looked identical) but the newer one was half the speed of the old one! I contacted Lexar to ask if it was a fake and they said it was not a fake and 'we don't give any spec for these drives and reserve the right to change them' - or words to that affect!
So just because you benchmark one UFD now does not mean that someone else's 'identical model' UFD will give the same performance!
#29
Posted 22 June 2010 - 06:04 PM
Funny that you mention Lexar, as this was the company i was contacting too.I had a similar issue with two 2GB Lexar Firefly Jumpdrives.
#30
Posted 11 July 2010 - 01:52 PM
Many UFD-manufacturers will change one or both without notice, will use different components for
UFD's with the same model-name but different capacities etc etc ..
Unfortunately the only nearly 100% certain way to identify the parts used is to disassemble the UFD .
BTW : Some of those IOPS-results are just horrible ..
In my experience that is more important than raw read/write speeds if you plan on running a full OS
from the UFD , ie, mediocre read/write speeds, but good IOPS-performance is better than good read/write but lousy IOPS ..
Another often overlooked thing is that UFD's with SLC-NAND are a way better choice for hosting a OS as they generally give better real-life performance and for a longer time to ..
#31
Posted 11 July 2010 - 02:11 PM
Besides SLC based Sticks being way more expansive and belonging to a dying species, as most manufacturers are switching to MLC even for their top performance line.Another often overlooked thing is that UFD's with SLC-NAND are a way better choice for hosting a OS as they generally give better real-life performance and for a longer time to ..
What you say is only true, for running a OS straight of the Stick, probably complete with alowing the OS to write data to it.
Most people around here use RamBoot methodes, which only use the Stick to read one 'large' disk image into RAM upon boot.
For this scenario, SLC based sticks do not present any advantage. Max. read speed, for a file measuring above 100MB, is the only thing that counts.
#32
Posted 11 July 2010 - 03:05 PM
Unfortunately the only nearly 100% certain way to identify the parts used is to disassemble the UFD .
Me thinks that Chipgenius accuracy is enough when compared to actually needing to disassemble the thingy....
http://www.boot-land...?showtopic=4661
Wonko
#33
Posted 27 July 2010 - 06:06 AM
-----------------------------------------------------------------------CrystalDiskMark 3.0 © 2007-2010 hiyohiyo Crystal Dew World : [url="http://crystalmark.info/"]http://crystalmark.info/[/url]-----------------------------------------------------------------------* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s] Sequential Read : 30.624 MB/s Sequential Write : 13.361 MB/s Random Read 512KB : 30.477 MB/s Random Write 512KB : 1.871 MB/s Random Read 4KB (QD=1) : 7.140 MB/s [ 1743.1 IOPS] Random Write 4KB (QD=1) : 0.021 MB/s [ 5.2 IOPS] Random Read 4KB (QD=32) : 8.069 MB/s [ 1970.0 IOPS] Random Write 4KB (QD=32) : 0.033 MB/s [ 8.0 IOPS] Test : 100 MB [E: 57.9% (8.3/14.3 GB)] (x1) Date : 2010/07/27 10:02:39 OS : Windows XP Professional SP3 [5.1 Build 2600] (x86)
#34
Posted 27 July 2010 - 10:31 AM
Really nifty design!
As you can see, it handles big files really well.
Interestingly, while the write performance is, even with small chunks, better than with its cheaper cousin, the read performance for small chunks is slightly worst.
#35
Posted 28 July 2010 - 01:15 PM
I have found other differences on two 4GB Lexar FireFlies, bought from the same shop (Staples) within a few days. The BootIt bit-flip does not work on one of them. I bought an 8GB Firefly at the same time. It works on that - as well as on two 16GB Fireflies. I used USBDeview.exe to inspect them. The one that is not bootable has a different device name - "USB Flash Drive" rather than "JD FireFly". On the cases of the 4GB drives, the bootable one is Rev AD and the unbootable one is Rev AJ. The cases are more or less identical, but I note some slight differences in the moulding of the word Lexar.I had a similar issue with two 2GB Lexar Firefly Jumpdrives. Supposedly identical models (they looked identical) but the newer one was half the speed of the old one! I contacted Lexar to ask if it was a fake and they said it was not a fake and 'we don't give any spec for these drives and reserve the right to change them' - or words to that affect!
I am going back to the shop to rummage through their bin of 4GB Fireflies to see if I can visually match the bootable one.
#36
Posted 28 July 2010 - 03:10 PM
I have found other differences on two 4GB Lexar FireFlies, bought from the same shop (Staples) within a few days. The BootIt bit-flip does not work on one of them. I bought an 8GB Firefly at the same time. It works on that - as well as on two 16GB Fireflies. I used USBDeview.exe to inspect them. The one that is not bootable has a different device name - "USB Flash Drive" rather than "JD FireFly". On the cases of the 4GB drives, the bootable one is Rev AD and the unbootable one is Rev AJ. The cases are more or less identical, but I note some slight differences in the moulding of the word Lexar.
I am going back to the shop to rummage through their bin of 4GB Fireflies to see if I can visually match the bootable one.
Use ChipGenius *instead* of USBDeview.exe:
http://www.boot-land...?showtopic=4661
It is very probable that it can identify the chip and consequently we may be able to find the specific tool to "flip the bit"-
Wonko
#37
Posted 28 July 2010 - 03:24 PM
Thanks, but that did not identify the chip on either bootable or non-bootable drive ("No match record"). The difference between them is only in the Device ID and the Serial Number:Use ChipGenius *instead* of USBDeview.exe. It is very probable that it can identify the chip and consequently we may be able to find the specific tool to "flip the bit"
Bootable:
PnP Device ID: VID = 05DC PID = A761
Serial Number: 5ZCGZ1TJEOO5BQUX0DCY
Non-bootable:
PnP Device ID: VID = 05DC PID = A810
Serial Number: AAJFDAUROYSC1GJ8
I note that all the other bootable Lexar FireFlies have a PID A7xx whereas the non-bootables have a PID of A8xx.
#38
Posted 28 July 2010 - 03:50 PM
definitely uses an SMI chip (and Chipgenius *should* be able to detect them. )PnP Device ID: VID = 05DC PID = A761
http://flashboot.ru/iflash.html
The:
is seemingly unknown (yet )PnP Device ID: VID = 05DC PID = A810
But what do you meean by "bootable" and "not -bootable" the difference should only be "flippable bit" with Lexar Utility or "not-flippable bit" with it.
Just for the record, the serial number means nothing, you can (with the appropriate utility) set it to whatever you want.
Traditionally Lexar uses SMI chips, but maybe this A810 is a newish/latest something.
Wonko
#39
Posted 28 July 2010 - 04:09 PM
Yes, that's what I mean. I was going to say "make-bootable-able." I have also tried several other utilities to sort out this flippin' bit - not just BootIt.But what do you mean by "bootable" and "not -bootable" the difference should only be "flippable bit" with Lexar Utility or "not-flippable bit" with it.
#40
Posted 28 July 2010 - 05:33 PM
I have just been back to the shop and found two flippable 4GB Lexar FireFlies at the back of the display. The only way I could identify them without removing the packaging was by the indented word "Lexar". The older, flippable, ones have very slightly thinner text. These ones will flip with BootIt. The later, unflippable, ones will not. I took my two example drives in to compare because the differences are so slight.I am going back to the shop to rummage through their bin of 4GB Fireflies to see if I can visually match the bootable one.
#41
Posted 02 August 2010 - 09:24 AM
I brought two JD Firefly USB 4bg sticks from staples. I need to use them for my work to boot up and build machines. First worked great. Second didn't. I've wasted so much time reformatting etc. etc. Took the "faulty" one back to Staples last weekend and had it replaced - same problem
As you say I notice the text between the two sticks are very slightly different - also the "good" one is seen as a JD Jumpfly whereas the other simply lists as removable disk in Win 7
I wonder whether there's a firmware upgrade or anything - else I'll have to take it back and just buy a different model
#42
Posted 02 August 2010 - 09:34 AM
I run WinXP, so I have to use a separate utility to make usb drives bootable. I understand Win7 has a built-in utility to make them bootable. Is this what you are using?As you say I notice the text between the two sticks are very slightly different - also the "good" one is seen as a JD Jumpfly whereas the other simply lists as removable disk in Win 7
The price of these 4GB Fireflies is good at £6.99 - maybe too good. Perhaps some of them are from a dodgy source?
#43
Posted 02 August 2010 - 09:39 AM
1. Diskpart
2. List disk and check EXACTLY the correct disk number
3. Select disk 1 (or whatever)
4. Clean
5. Create partition primary
6. Select partition 1
7. Active
8. Format fs=fat32 quick
9. Assign
10. Exit
I can only use Windows 7 for this since XP doesn't see the drive at all. AS I said one works great the other doesn't - I thought I was going mad until I saw your thread - its so annoying that supposedly identical products are actually different
I don't want to keep trawling back to Staples so I might get in touch with Lexar to see if they have a solution - if you have any thoughts let me know
... one thought I did have was that it might need Vista - I haven't got a copy of Vista in front of me but that still doesn't excuse one working fine under win7 and the other not
#44
Posted 02 August 2010 - 09:53 AM
Shouldn't there be a bootsect.exe here somewhere?I'm using diskpart to list and format the disk
More than one person can be mad at the same timeI thought I was going mad until I saw your thread
I tried a friend's Vista, but it did not have all the commands I was looking for, and diskpart did not list this drive (although it was visible in Explorer)... one thought I did have was that it might need Vista - I haven't got a copy of Vista in front of me but that still doesn't excuse one working fine under win7 and the other not
#45
Posted 02 August 2010 - 03:15 PM
I have sent email to Lexar support and received nondescript answer on how to format a USB device !!!
In the end I'll just go back to Staples junk it an get another type of USB drive - perhaps a Kingston or HP or something - just didn't want to have to make a special trip again
#46
Posted 09 August 2010 - 12:18 PM
Over the weekend I return the Lexar stick and purchased a Sandisk Cruzar which was £12 - worked perfectly first time !
I guess you get what you pay for
#47
Posted 09 August 2010 - 01:21 PM
Spec: 30 MB/s (read) 24 MB/s (write)
#48
Posted 10 August 2010 - 11:12 AM
Has anyone tried the new Intenso High Speed Line USB sticks yet? £45 for 16GB.
Spec: 30 MB/s (read) 24 MB/s (write)
Some groundwork can be found at Nirsoft. But who knows what you get at home, putting your hands on it in your favorite file manager. As the time goes by vendor can simplify design -- it can be spoiled by the cheap memory, not dual channel, another controller, etc.
I own two (well, plus PQI U330 with ~15Mb/s linear write) realy fast flash drives Transcend branded: JetFlash 168 (the fastest series from Transcend, SLC -- phase out) 2Gb, and JetFlash 600 16Gb. JF168 is very fast, even on small blocks, but only 2Gb, so I bought JF600 16Gb for my GRUB-based recovery media (a lof of PEs, Linux, Win setup, etc).
Transcend JF600 16Gb
---------------------------------------------------------------
CrystalDiskMark 3.0 Beta1 © 2007-2009 hiyohiyo
Crystal Dew World : http://crystalmark.info/
---------------------------------------------------------------
Sequential Read : 34.982 MB/s
Sequential Write : 19.001 MB/s
Random Read 512KB : 34.139 MB/s
Random Write 512KB : 1.470 MB/s
Random Read 4KB (QD=1) : 4.970 MB/s [ 1213.3 IOPS]
Random Write 4KB (QD=1) : 0.014 MB/s [ 3.4 IOPS]
Random Read 4KB (QD=32) : 5.042 MB/s [ 1230.9 IOPS]
Random Write 4KB (QD=32) : 0.027 MB/s [ 6.6 IOPS]
Test : 100 MB [H: Used 0.0% (0.0/15.0 GB)]
Date : 2010/06/04 20:24:06
OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x86)
Transcend JF600 16Gb also, but 1Gb test file use:
---------------------------------------------------------------
CrystalDiskMark 3.0 Beta1 © 2007-2009 hiyohiyo
Crystal Dew World : http://crystalmark.info/
---------------------------------------------------------------
Sequential Read : 36.541 MB/s
Sequential Write : 21.314 MB/s
Random Read 512KB : 35.718 MB/s
Random Write 512KB : 1.399 MB/s
Random Read 4KB (QD=1) : 4.948 MB/s [ 1207.9 IOPS]
Random Write 4KB (QD=1) : 0.010 MB/s [ 2.5 IOPS]
Random Read 4KB (QD=32) : 5.012 MB/s [ 1223.7 IOPS]
Random Write 4KB (QD=32) : 0.011 MB/s [ 2.6 IOPS]
Test : 1000 MB [H: Used 0.0% (0.0/15.0 GB)]
Date : 2010/06/04 20:31:52
OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x86)
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users