SSD Write Endurance 25nm Vs 34nm

[Hopalong X]

If that is true, then a SandForce 2281 is a full 25 percent more fun than non-SF2281's!

Man, SandForce should hire me for their marketing... now with 25 percent more "Fun" in every box!

"It's so easy that even a five year old can do it." Note-1

They didn't ship one in the box?????

Note-1: 5 year old not inclucded.

[Christopher]

*Taiwanese firmware engineer sold separately.

[Hopalong X]

*Taiwanese firmware engineer sold separately.

If you had one of those it would be fixed!

[zads]

When SSD Write Endurance testing is part of your job,

This is for one 25nm test:



(chamber temperature is 0-75C)

[Christopher]

Mushkin Chronos Deluxe 60 Update, Day 33
Updated to 3.3.0 FW

05 2
Retired Block Count

B1 13
Wear Range Delta

F1 312945
Host Writes

E9 241375
NAND Writes

E6 100
Life Curve

E7 10
Life Left

Average 127.47MB/s Avg
Intel RST drivers, Celeron G530 Biostar TH67+

755 Hours Work (24hrs since the last update)
Time 1 month 1 day 11 hours

9 GiB Minimum Free Space


I'm still waiting on 3.3.2 FW, but no sign yet. I'd imagine it should be around in a week or two.


??
If I can tame the file delete problems, I can add or remove static data at will. Right now I have 12.7GB of static data. There has never been more than 17 and never less than 12GB.

12.7GB/55.7 is 22.8 percent static data.

[Anvil]

When SSD Write Endurance testing is part of your job,

This is for one 25nm test:
...
(chamber temperature is 0-75C)

Looks expensive
What's the workload?

--

Kingston SSDNow 40GB (X25-V)

421.75GB Host writes
Reallocated sectors : 12

MD5 OK

34.31MiB/s on avg (~35 hours)

--

Corsair Force 3 120GB

01 95/50 (Raw read error rate)
05 2 (Retired Block count)
B1 87 (Wear range delta)
E6 100 (Life curve status)
E7 38 (SSD Life left)
E9 235195 (Raw writes)
F1 313175 (Host writes)

MD5 OK

102.61MiB/s on avg (~35 hours)

power on hours : 940

[Hopalong X]

Looks expensive
What's the workload?

Ice-T and Ice Cube iTunes.

[One_Hertz]

7257 reallocations. 531.4TB. Max speed down to 37.5MB/s (!).

I am shutting off my rig for a few days to see if it can hold the data in tact.

[bluestang]

M225->Vertex Turbo 64GB Update:

581.73 TiB (639.62 TB) total
1565.52 hours
11705 Raw Wear
117.98 MB/s avg for the last 16.32 hours (on W7 x64)
MD5 OK
C4-Erase Failure Block Count (Realloc Sectors) at 10.
(1=Bank 6/Block 2406; 2=Bank 3/Block 3925; 3=Bank 0/Block 1766; 4=Bank 0/Block 829; 5=Bank 4/Block 3191; 6=Bank 7/Block 937; 7=Bank 7/Block 1980; 8=Bank 7/Block 442; 9=Bank 7/Block 700; 10=Bank 2/Block 1066)

[Christopher]

3.3.2 just released for the MCDx.

I just upgraded so I can see what happens...

I'm going to leave it as is until Thursday, then I'm going to put it back in my main system over the weekend and see what happens. If its still working when I get back on Sunday, then I think it will be looking good.

[B.A.T]

Todays update:
m4
751.8592 TiB
2687 hours
Avg speed 88.72 MiB/s.
AD gone from 182 to 179.
P/E 13019.
Value 01 (raw read error rate) has changed from 16 to 23.
MD5 OK.
Reallocated sectors : 00


Kingston V+100:
172.8184 TiB
683 hours
Avg speed 76.11 MiB/s
AD gone from 26 to 20.
P/E?
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
7566.5GB Host writes
18652 hours
Reallocated sectors : 00
MWI=92
MD5 OK
59.56MiB/s on avg


I'm still working on the MTRON

[Christopher]

I'm seeing a small but noticeable decrease in speed in terms of average MB/s with the new 3.3.2 FW.

The Samsung 830 launched in the States today and is generally available. Supposedly, it comes with Norton Ghost AND a voucher for Batman: Arkham Asylum, according to thessdreview.com, but I don't see it listed as being included with the retail kit at Newegg.com.

The 64GB version might be a good test candidate...

[B.A.T]

I agree. The Samsung 830 is a good candidate

[Christopher]

I've heard that Samsung says that all of the capacities of 830 perform the same... while I'm not sure I believe that (the 64 can't be as fast as the 256). If it's just a much faster 470, you could kill the 470 in about 10 days...

I'm sure that no review site will get a 120GB 830, but its probably almost as fast as the 256/512s. The 830 has some pretty good measured latency and astounding sequential writes.

[B.A.T]

According to Anandtech it's only the read speed that is the same on all capacities. Write decreases with the smaller sizes.

[Christopher]

That makes a lot more sense. 520MBs reads and 160MBs on writes for the 64GB model according to Anand's chart. That is actually pretty good. If you test the Mushkin with incompressible data you get about those figures. With the Mushkin though, if you ever get the urge to write a couple TB of zeros, you can do so at 300MBs sustained. Why you would ever want to, I don't know, but it's nice to know you could. The 64GB 830 would seem to be a good deal faster than the M4 64GB, but I suspect those two might be pretty close in actual performance.

[B.A.T]

Well, we'll just have to wait and see. It depends on the controllers ability to handle the load, but maybe it's just as good or better than the m4.

[Christopher]

What's going on with the C300?

[B.A.T]

I don't know. Haven't seen anything from Vapor in a long time.

[Christopher]

Well, it's probably time for a chart update too.

[B.A.T]

It's been 200TiB since the last update

[Anvil]

Kingston SSDNow 40GB (X25-V)

423.28GB Host writes
Reallocated sectors : 12

MD5 OK

33.86MiB/s on avg (~49 hours)

--

Corsair Force 3 120GB

01 89/50 (Raw read error rate)
05 2 (Retired Block count)
B1 86 (Wear range delta)
E6 100 (Life curve status)
E7 37 (SSD Life left)
E9 238892 (Raw writes)
F1 318081 (Host writes)

MD5 OK

102.60MiB/s on avg (~49 hours)

power on hours : 954

B1 (WRD) has gone from 87 to 86, looks like there is some activity on wear leveling

[Christopher]

Mushkin Chronos Deluxe 60 Update, Day 34
3.3.2FW

05 2
Retired Block Count

B1 13
Wear Range Delta

F1 323050
Host Writes

E9 249173
NAND Writes

E6 100
Life Curve

E7 10
Life Left

Average 126.91MB/s Avg
Intel RST drivers, Celeron G530 Biostar TH67+

778 Hours Work (24hrs since the last update)
Time 2 month 10 hours

11 GiB Minimum Free Space




I missed the 2 new reallocation events today. In the grand scheme of things two reallocs aren't very consequential, but I think it's noteworthy until it happens every day. The M4 is smoking the competition in that respect.

Speed went back up to normal, about 129MBs.

[Anvil]

Kingston SSDNow 40GB (X25-V)

424.52GB Host writes
Reallocated sectors : 12

MD5 OK

33.66MiB/s on avg (~60 hours)

--

Corsair Force 3 120GB

01 92/50 (Raw read error rate)
05 2 (Retired Block count)
B1 85 (Wear range delta)
E6 100 (Life curve status)
E7 37 (SSD Life left)
E9 241890 (Raw writes)
F1 322068 (Host writes)

MD5 OK

102.56MiB/s on avg (~60 hours)

power on hours : 965

B1 (WRD) is down again, from 86 to 85, slowly decreasing.

[bluestang]

M225->Vertex Turbo 64GB Update:

588.11 TiB (646.64 TB) total
1579.88 hours
11822 Raw Wear
117.46 MB/s avg for the last 15.76 hours (on W7 x64)
MD5 OK
C4-Erase Failure Block Count (Realloc Sectors) at 10.
(1=Bank 6/Block 2406; 2=Bank 3/Block 3925; 3=Bank 0/Block 1766; 4=Bank 0/Block 829; 5=Bank 4/Block 3191; 6=Bank 7/Block 937; 7=Bank 7/Block 1980; 8=Bank 7/Block 442; 9=Bank 7/Block 700; 10=Bank 2/Block 1066)