Sun Oracle F40 400GB PCI-e SSD - 65$

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
I don't have one, but I've been considering one for a while. One of the attractions about this is that it's not proprietary; it shows up as a LSI controller with 4x SSDs attached to it. Many OSes already natively have the driver for the LSI controller so it's easy to get going; and I think this also means you can boot off this card. I never got one of these, but I did get a few Fusion-IO cards, which are great in their own right, but they require special drivers, and hence also can't boot.

this is very tempting... but i'm telling myself I'll hold off unless the buyer accepts $50/each.
When you think about it the latency on these is probably amazing, since the storage is so close to the controller electrically wise.
 

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
Haven't found that in any documentation but they have eMLC and quite a significant amount of overprovisioning. I'd hazard a guess of 10DWPD for five years.

They are however not the most reliable in my experience so Patrick's suggestion of spares is very apt. The good part is that you can just replace an individual drive (there are 4x 100G) on the card so one spare card can go a long way. The bad is that based on how many people were complaining in prior threads it's not unlikely you'll need to replace parts sooner rather than later (at least compared to other drives).
So what you're saying is, using 4 of these in a striped config is probably not the safest course of action? If modules fail as often as you suggest you basically have a whole bunch of volatile raid 0's on hand.
 

mmk

Active Member
Oct 15, 2016
104
39
28
Czech Republic
So what you're saying is, using 4 of these in a striped config is probably not the safest course of action? If modules fail as often as you suggest you basically have a whole bunch of volatile raid 0's on hand.
I'd only ever use RAID0 for irrelevant or temporary data. For home use one of these wouldn't be a too bad setup for a ZFS L2ARC/SLOG or some other equivalent purpose. For work I'd grab two in order not to waste time.
 

mmk

Active Member
Oct 15, 2016
104
39
28
Czech Republic
Received mine today. Out of the three I ordered, one had lost a capacitor on the way. The same thing happened last time I ordered a bunch of these cards (from a different seller). Argh. A couple more minutes of packaging would have fixed this..

For the good part: all the modules have 100% of ssd life left in them! They do have a few years of runtime on them but must have been rather quiet on the IO front..
 
  • Like
Reactions: frogtech

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
Received mine today. Out of the three I ordered, one had lost a capacitor on the way. The same thing happened last time I ordered a bunch of these cards (from a different seller). Argh. A couple more minutes of packaging would have fixed this..

For the good part: all the modules have 100% of ssd life left in them! They do have a few years of runtime on them but must have been rather quiet on the IO front..
Can you post some quick crystaldiskbench numbers from a single module and then some numbers from all 4 striped?
 

Marsh

Moderator
May 12, 2013
2,644
1,496
113
From my Aug 2016 notes

4 x 93gb SSD

Tyan S5533 ITX
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1913.292 MB/s
Sequential Write (Q= 32,T= 1) : 639.494 MB/s
Random Read 4KiB (Q= 32,T= 1) : 355.791 MB/s [ 86863.0 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 163.149 MB/s [ 39831.3 IOPS]
Sequential Read (T= 1) : 1636.540 MB/s
Sequential Write (T= 1) : 604.922 MB/s
Random Read 4KiB (Q= 1,T= 1) : 26.764 MB/s [ 6534.2 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 33.653 MB/s [ 8216.1 IOPS]

Test : 8192 MiB [E: 0.0% (0.1/372.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/25 10:41:41
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Tyan S5533 ITX
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1926.593 MB/s
Sequential Write (Q= 32,T= 1) : 544.163 MB/s
Random Read 4KiB (Q= 32,T= 1) : 347.484 MB/s [ 84835.0 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 151.896 MB/s [ 37084.0 IOPS]
Sequential Read (T= 1) : 1606.786 MB/s
Sequential Write (T= 1) : 496.208 MB/s
Random Read 4KiB (Q= 1,T= 1) : 22.864 MB/s [ 5582.0 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 33.632 MB/s [ 8210.9 IOPS]

Test : 8192 MiB [E: 0.0% (0.1/372.0 GiB)] (x3) [Interval=5 sec]
Date : 2016/08/25 10:22:43
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Intel X79 E5-1607
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1528.141 MB/s
Sequential Write (Q= 32,T= 1) : 703.699 MB/s
Random Read 4KiB (Q= 32,T= 1) : 561.244 MB/s [137022.5 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 176.489 MB/s [ 43088.1 IOPS]
Sequential Read (T= 1) : 1351.620 MB/s
Sequential Write (T= 1) : 636.613 MB/s
Random Read 4KiB (Q= 1,T= 1) : 29.968 MB/s [ 7316.4 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 27.604 MB/s [ 6739.3 IOPS]

Test : 1024 MiB [E: 0.0% (0.1/372.0 GiB)] (x3) [Interval=5 sec]
Date : 2016/08/25 9:50:02
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Intel X79 E5-1607
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1514.853 MB/s
Sequential Write (Q= 32,T= 1) : 715.073 MB/s
Random Read 4KiB (Q= 32,T= 1) : 461.758 MB/s [112733.9 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 172.188 MB/s [ 42038.1 IOPS]
Sequential Read (T= 1) : 1337.167 MB/s
Sequential Write (T= 1) : 582.384 MB/s
Random Read 4KiB (Q= 1,T= 1) : 23.800 MB/s [ 5810.5 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 33.288 MB/s [ 8127.0 IOPS]

Test : 4096 MiB [E: 0.0% (0.1/372.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/25 9:07:12
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Intel X79 E5-1607
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1498.026 MB/s
Sequential Write (Q= 32,T= 1) : 625.191 MB/s
Random Read 4KiB (Q= 32,T= 1) : 334.814 MB/s [ 81741.7 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 158.300 MB/s [ 38647.5 IOPS]
Sequential Read (T= 1) : 1346.681 MB/s
Sequential Write (T= 1) : 580.159 MB/s
Random Read 4KiB (Q= 1,T= 1) : 23.426 MB/s [ 5719.2 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 31.905 MB/s [ 7789.3 IOPS]

Test : 8192 MiB [E: 0.0% (0.1/372.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/25 9:01:18
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Tyan S5533
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 2003.832 MB/s
Sequential Write (Q= 32,T= 1) : 637.255 MB/s
Random Read 4KiB (Q= 32,T= 1) : 475.095 MB/s [115990.0 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 171.273 MB/s [ 41814.7 IOPS]
Sequential Read (T= 1) : 1623.941 MB/s
Sequential Write (T= 1) : 622.874 MB/s
Random Read 4KiB (Q= 1,T= 1) : 27.131 MB/s [ 6623.8 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 30.844 MB/s [ 7530.3 IOPS]

Test : 4096 MiB [E: 0.0% (0.1/372.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/24 13:11:40
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)

-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 1907.978 MB/s
Sequential Write (Q= 32,T= 1) : 660.281 MB/s
Random Read 4KiB (Q= 32,T= 1) : 346.657 MB/s [ 84633.1 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 159.046 MB/s [ 38829.6 IOPS]
Sequential Read (T= 1) : 1639.774 MB/s
Sequential Write (T= 1) : 569.117 MB/s
Random Read 4KiB (Q= 1,T= 1) : 26.832 MB/s [ 6550.8 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 36.522 MB/s [ 8916.5 IOPS]

Test : 8192 MiB [E: 0.0% (0.1/372.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/25 17:06:28
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)


Single SSD
-----------------------------------------------------------------------
CrystalDiskMark 5.1.2 x64 (C) 2007-2016 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

Sequential Read (Q= 32,T= 1) : 536.602 MB/s
Sequential Write (Q= 32,T= 1) : 115.915 MB/s
Random Read 4KiB (Q= 32,T= 1) : 89.915 MB/s [ 21951.9 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 58.506 MB/s [ 14283.7 IOPS]
Sequential Read (T= 1) : 509.141 MB/s
Sequential Write (T= 1) : 166.297 MB/s
Random Read 4KiB (Q= 1,T= 1) : 21.450 MB/s [ 5236.8 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 25.243 MB/s [ 6162.8 IOPS]

Test : 4096 MiB [E: 0.1% (0.1/93.0 GiB)] (x5) [Interval=5 sec]
Date : 2016/08/24 12:41:03
OS : Windows Server 2012 R2 Datacenter (Full installation) [6.3 Build 9600] (x64)
 

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
Got mine the other day, finally testing them. Most of my drives have either 24 TB or 78 TB (~24,000 or ~77/78,000) written, most are 100% health with a small number showing as 99% health in crystaldiskinfo even though their GB written is lower. All roughly around 28,000 hours POT, with 73-76~ power on count.

Can anyone report what their numbers are? I suspect they'll be roughly the same.
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
Display WarpDrive Health >

Code:
Retired Block Count                   0
Power-On Hours                        17951.4
Device Power Cycle Count              48
Gigabytes Erased                      72074      (Gigabytes)
Reserved (over-provisioned) Blocks    35008
Program Fail Count                    0
Erase Fail Count 1                    0
Unexpected Power Loss Count           63
I/O Error Detection Code Rate         0
Uncorrectable RAISE Errors            0
Maximum Lifetime Temperature          50         (degree C)
Cached SMART Data Age                 00:00:00   (Hours:Minutes:Seconds)
SSD Life Left (PE Cycles)             100        (%)
Total Writes From Host                121809
Total Reads To Host                   833202
Write Amplification                   0.57
Reserved Blocks Remaining             100        (%)
Warranty Remaining                    100        (%)
Code:
Retired Block Count                   0
Power-On Hours                        32353.7
Device Power Cycle Count              105
Gigabytes Erased                      113969     (Gigabytes)
Reserved (over-provisioned) Blocks    35200
Program Fail Count                    0
Erase Fail Count 1                    0
Unexpected Power Loss Count           119
I/O Error Detection Code Rate         0
Uncorrectable RAISE Errors            0
Maximum Lifetime Temperature          49         (degree C)
Cached SMART Data Age                 00:00:00   (Hours:Minutes:Seconds)
SSD Life Left (PE Cycles)             100        (%)
Total Writes From Host                191937
Total Reads To Host                   971678
Write Amplification                   0.57
Reserved Blocks Remaining             100        (%)
Warranty Remaining                    100        (%)
 

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
So I bought some spares from a member on here (they work fine btw) that I wanted to use in my build because the member said they were all %100 health (which they were).

During my progress of getting my other 8 all up to 100% by switching out modules, can anyone tell me what kind of a condition would have to occur for the drive healths of these things to drop to 99% from 100%? Basically what I am doing is looking through all of the adapters for 99% health modules, and then making notes of serial numbers, disassembling, and swapping out until I have 8 adapters that are completely 100% across the board. This bugs me because the ones with 99% health do not have anything different about them in Crystaldiskinfo that indicates any reason why the health percentage dropped...

Any thoughts?
 
  • Like
Reactions: Monoman

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
Another follow up on this...swapping out the modules is super annoying. I stripped screws on 2 of them because they're so ridiculously over-torqued. And then re-installing the pair of disks onto the ribbon module that connects the storage to the controller is kind of awkward. Not the easiest things ever. I would say there's a good chance I'd avoid these in the future or at least buy just one to dick around with, not 8 or more like I had.
 

fishtacos

New Member
Jun 8, 2017
23
13
3
Another follow up on this...swapping out the modules is super annoying. I stripped screws on 2 of them because they're so ridiculously over-torqued. And then re-installing the pair of disks onto the ribbon module that connects the storage to the controller is kind of awkward. Not the easiest things ever. I would say there's a good chance I'd avoid these in the future or at least buy just one to dick around with, not 8 or more like I had.
So you decided to switch out parts of the 4 modules per card to get full 100% drive health on all of them, for no reason other than OCD, then complained about stripping the screws on 2 of them?
 

frogtech

Well-Known Member
Jan 4, 2016
1,482
272
83
35
So you decided to switch out parts of the 4 modules per card to get full 100% drive health on all of them, for no reason other than OCD, then complained about stripping the screws on 2 of them?
Uhhh, yeah? Lol. I mean I plan on using them in a converged manner. So I want the 8 in use to be as in relatively top shape as possible. I don't think this is unreasonable, really, if you consider past posts from other members who have complained about modules randomly dying. In a way, if those 99 percenters were at some stage of fault, I did myself a favor pre-emptively by swapping them out an ensuring all 32 modules across 8 devices are at 100%. This is pure speculation but I'm okay with it.
 

fishtacos

New Member
Jun 8, 2017
23
13
3
Uhhh, yeah? Lol. I mean I plan on using them in a converged manner. So I want the 8 in use to be as in relatively top shape as possible. I don't think this is unreasonable, really, if you consider past posts from other members who have complained about modules randomly dying. In a way, if those 99 percenters were at some stage of fault, I did myself a favor pre-emptively by swapping them out an ensuring all 32 modules across 8 devices are at 100%. This is pure speculation but I'm okay with it.
No, not at all unreasonable. 99% drive health vs 100% is a tremendous indicator for the work that you put in, and later complained about, and are now still defending.