SSD Thread (Benchmarks, Brands, News, and Advice)

Does anybody have a copy of the Mushkin Maindiag.exe program for SSD firmware updates?

I tried the different links by 5150Joker, but no go.

Please help!

 

I have something but I'm not sure if this is the right thing.

 

Tomy B., he's cross-posted all over the place and 'solved' it himself.

Doesn't even share...

 

Here are Anandtech thoughts on the subject,

The LSI SandForce Acquisition: Anand's Thoughts

 

Can anyone point to me benches/reviews comparing old Intels (first and second gen, SLC and MLC x-25E and x-25M) to the new ones currently available? I'm very curious to read how they compare and how they've improved aside from fitting more nand in there.

 

Have you seen this?

See:
Intel SSD 710 Tested: MLC NAND Flash Hits The Enterprise : Intel On Enterprise Storage: No More SLC; Meet HET MLC

 

That review was done with old firmware. Here's how the Crucial M4 performs with fw009.

Crucial m4 | Trace-based benchmarks | Core | Tweakers.net Reviews

Saying that Intel 510 is more reliable than Crucial M4 is only based on assumptions. Crucial has way more experience with the Marvell controller than Intel.

It's extremely unlikely that anyone would notice a difference between Intel 510 and Crucial M4 in performance or reliability. In other words buy whichever is cheaper.

 

Phil, I'm noting that the higher 4K R r/w's of the M4 could, in a specific scenario, put it ahead for certain users.

In a workstation type use/load (ie: mine) the 510 is ahead overall - sequential speeds are not something to scoff at offhand.

As for the 510 being more reliable being based on assumptions - well, maybe, but this post doesn't help the M4's rep any (for me):

See:
http://forum.notebookreview.com/8043180-post584.html


The 'StorageMark 2011' scores don't prove anything to me (even if I could read German?). At the most they indicate what I've been saying: if high 4K R r/w's are a large part of your workload, the M4 could be the better fit. (This is what I'm assuming another synthetic benchmark is measuring/rating once again).

 

+1 to that....

I got M4 128GB for 195$ while intel 510 120GB was selling at 315$....
so it was no brainer for me....

 

Please stop posting misleading post like this. In workstation load the M4 128GB clearly beats the Intel 510 120GB by more than a 10% margin as can be seen by the Tweakers.net benchmarks (the 250GB Intel performs better). Unlike Anandtechs benchmarks the Tweakers benchmarks are real time AND real world.

I guess you meant that for your specific usage (large file transfers) the Intel is the better choice. That may be true although I'm not sure about that. Especially the 128GB M4 with firmware 009 I don't see loosing out to the Intel 120GB.

And like I said, it's highly unlikely anyone will notice any difference in real life scenarios even when transferring files.

So you found a negative experience of the M4. I've seen several negative experiences with Intel 510. Does that make the 510 an unreliable drive? Of course not.

Intel messed up the firmware for SSDs based on their own controller several times. Does that make them an unreliable company? No again.

Reliability for SSD drives can only be judged from large samples. There is zero evidence to support that the M4 is not a reliable drive.

 

Why is my direct experience 'misleading'? When I'm clearly labeling/stating it as such?

I'd be curious to see a link to those misbehaving 510's?

And for the record: I don't use/trust SSD's without fully populated controller channels.

160GB or larger are the only 'sane' reasons to go to an SSD for me. I don't buy them for their namesake, I need them for the performance they offer.

Anything smaller in capacity (like those in the tests linked to) are just not serious enough for my uses nor my specific setup requirements.

Does that site show their real world traces with the larger SSD's tested?

Maybe you could request them to test one with my setups: use only 100GB out of 250GB (not simply 'unused' capacity, but specifically 'unallocated' capacity via partitioning)?

One last quick question: how long do those traces last for? Just curious.

 

Agreed. Most of the times the Intel is much more expensive.

Because you didn't mention that your workload is a unique workload: transferring many large files. That is not what is commonly considered as typical workstation type workload.

Ok I will play that game once: http://gathering.tweakers.net/forum/view_message/36301538. And that is not the only one I've seen. Especially in combination with Mac there have been huge problems.

But as I said, linking to individual bad experiences has no statistical significance.

Yes. The 250GB is relatively a better performer. The Crucial M4 256GB performs similar to the 128GB.
Ssd-test: Memorights 240GB tegen Crucials 256GB | Trace-based benchmarks | Core | Tweakers.net Reviews

I don't know. Judging by his forum posts I do rate the guy who set it up as more knowledgeable than Anand.

Edit: by the way, the new Plextor M2P SSD is the new king of the hill in their setup.
http://tweakers.net/reviews/2365/3/ssd-test-plextor-pakt-de-kroon-trace-based-benchmarks.html

 

well....
this is getting kind of serious...........

 

It doesn't add to your reputation not recognizing that it's Dutch.

 

That's refreshing to hear as I have SF OCZ

Are these tweakers benchmarks in-house or available for download?

Well according to this review of Plextor PX-M2P SSD

Why is that? Is it the same Plextor drive or am I looking at something else?

 

The Plextor beats the Vertex 3 (and Maxiops) in all their real world benchmarks.

It's only the Kingston HyperX that wins some benchmarks, while the Plextor wins some others.

 

What do you think about RAIDing two different size drives 1.8" + 2.5"? If they're same capacity? I mean I guess there would be minute power differences that could affect performance differences and screw up the redundancy, especially if it's RAID-0, but I figured I'd ask if it'd been done before...

 

Combining something like two Intel X-25M's would be fine since they're the exact same drive whether they're in 1.8" formfactor or 2.5". Something with wildly different performance characteristics... I dunno.

 

Yeah, I was hoping the Swiss Techno Junkie would chime in 'cuz I know he uses 1.8" form factor. I don't know that he ever tried to RAID-0 it with something else though. But I DO need to RAID-0 because I'm trying to build a bigger volume with two small 32gb or 64gb drives.

 

I would say, try it and see what happens.

 

Phil, Thanks!

My workload is not simply transferring many large files; it also involves creating these huge files too... And by 'workstation' type workload, I mean one in which high 4K R r/w's do not make a discernable difference (vs. say a 'server' workload which thrives on high 4K R r/w's).

 

Hi guys, need a little help getting my SSD going in my Dell e4310. I had it installed in here once before after a few hickups but had since moved it to a Dell e6320. Didn't like the 6320 so I'm back to the 4310.

Here's where Im at so far. I installed the SSD and then went into the BIOS and set it to AHCI. I then inserted the Win7 disk and I dont remember if I did legacy boot or UEFI boot but I got win7 installed no problem. I then started updating drivers and everything was fine even after several reboots. Then all of a sudden this morning, I reboot and it boots directly into safe mode and is all sluggish and weird.

Any ideas? Am I missing something?

 

that would be me, then, right?

lets read up what the discussion is about, then..


yes, you should be fine (with similar/same ssds/hdds). i've used the 1.8" ssds in 2.5" systems with a simple adapter without problems. the sata part is identical, it's only the power part that's different (3.3V instead of 5V or something).

so while i never raided like that, i can tell you that the system does not behave different on 1.8" and 2.5" (or 3.5" for that matter), at least in sata world.

in pre-sata world, i have no clue about raiding, it mainly will depend on what the chipset can do.

 

Sure is


Thank you sir. That's what I was hoping, that the power difference doesn't make the striping act weird.

 

HI, Im a little new to the SSD, and with out knowing that defrag is bad for ssd i used game booster to defrag some games!! What kind of damage have i caused??

 

Just some additional writes, maybe shorten the life by a little bit, it is using it consistantly that will really shorten the life. You will be alright, just don't use defrag again. And there have been some forums that the users have used constant writing to see the actual lifetimes of ssd's, and they all (except for a SF drive that stopped really fast) outlived by a huge margin the expected lifetimes. So don't worry. See this thread: http://www.xtremesystems.org/forums/showthread.php?271063-SSD-Write-Endurance-25nm-Vs-34nm/page34

 

Someone is telling me that power consumption of SSDs would be 10 times less if you only enable DIPM.

Is this true? I don't have a laptop to test it on (only OS X now).

 

10x less of almost nothing is still 'almost nothing'.

Also, wouldn't you be essentially be disabling any background GC from happening?

Doesn't seem worth it (yet).

 

It seems DIPM is automatically enabled on laptops running Intel RST.

 

Beginning with the Intel® Rapid Storage Technology 10.0 release, LPM support is enabled by default on both mobile and desktop platforms.(from an official Intel document)
It also depends on the device firmware and the BIOS.

 

Yeah, 'automatically' has a lot of prerequisites as maximinimaus also stated.

I don't think any of my systems have it on - but I'll be able to check later tonight or tomorrow.

 

I'm interested how you would/could check it as a look in the registry is not sufficient.

 

I think the Intel RST utility has that capability.

 

What's that?

 

AHCI Link Power Management - Enable HIPM and DIPM - Windows 7 Forums

So DIPM = LPM ?

 

No, LPM = is a combination of DIPM + HIPM.
DIPM = Device Initiated Power Management
HIPM = Host Initiated Power Management.

The entry for AHCI LPM shown in the screenshot is not set by default, at least not for me.

For further information please read Intel document

http://www.intel.com/content/dam/doc/reference-guide/sata-devices-implementation-recommendations.pdf

 

No, I just checked it.

You can check the capabilities of the BIOS regarding the implementation for the RAID ROM with RaidCply.exe executed in a booted DOS environment.

Following is shown for my BIOS for port 0(configured for RAID0)

4.7 Port Implemented Bit Set for Port 0.......................ENABLED
4.8 PI Bit Matches PE Bit for Port 0.............................PASS
4.9 Aggressive Link Power Management on Port 0...............DISABLED
4.10 Hot Plug Support on Port 0..............................DISABLED
4.11 Interlock Switch Attached to Port 0..........................N/A
4.12 ALPE and HPCP Cannot Both be Set on Port 0..................PASS
4.13 ISP Requires HPCP on Port 0.................................PASS
4.14 Port 0 Command Register Locked..............................PASS
4.15 External Port Bit Set for Port 0........................DISABLED

 

Intel Bringing TRIM Support To RAID0 With RST 11.5

 

There is a dedicated thread for this topic:
http://forum.notebookreview.com/har...ades/625315-exciting-news-all-raid0-fans.html

 

OCZ has another card up their sleeve, a new Indilinx drive.



And low and behold... "OCZ is touting incompressible performance as a major advantage of the Octane over its Vertex 3/Agility 3 drives."

http://www.anandtech.com/show/5143/oczs-octane-initial-performance-results

 

hey guys, need some quick advice. i am stuck between choosing a 256 m4 and a 830. while the real world performance differences are negligible, i am more concerned about the power consumption. i have read on several forums that the new samsung 830 uses a lot less power than the crucial drive. some have even claimed up to a full hours difference on their laptops. does anyone that has these drives care to help me with this. i have been leaning towards the crucial drive, but if the samsung is that conservative on power, and still a little faster, i might need to rethink this. i have to make a decision today, or tomorrow at the very latest. thanks!

 

Are you sure they were comparing M4 to 830?

To my knowledge the Crucial uses less power and this is confirmed by Storagereview.
Samsung SSD 830 Review (256GB) | StorageReview.com
Crucial m4 SSD Review (256GB) | StorageReview.com

About the performance: I'd like to see Crucial M4 (firmware 009) compared to Samsung 830 in real world, real time benchmarks. I haven't seen them but I'd put my money on the M4. Not that the 830 is slow, certainly not.

 

ok, one last question phil, is the m4 better, faster and more reliable than the new c300?

 

It's faster yes no doubt.

More reliable? hard to say. I do think the M4 is the more reliable product but the C300 uses 34nm NAND, which can (should) last longer than 25nm NAND.

 

Most places don't even carry C300 anymore, shouldn't really be a factor for buying a new SSD. M4 has much better price points, but at cost with 25 nm flash NAND.

 

OK I can eat my words. Samsung 830 128GB beats Crucial M4 128GB firmware 009.
Ssd-test: Samsung probeert het opnieuw | Conclusie | Core | Tweakers.net Reviews

Not that one would notice the differences.

Power consumption: Crucial wins with a small margin.
Ssd-test: Samsung probeert het opnieuw | Energieverbruik | Core | Tweakers.net Reviews

Overall standings


And official Samsung lifetime figures.

 

Gaming?

Here is how an SSD will affect certain games with/without an SSD:

See:
Exploring SSD Performance In Battlefield 3, F1 2011, And Rift : Profiling Storage I/O In Three New Games

 

seems like a good test...

well, i liked snappiness of SSD though on gaming

 

what msata drive do you recommend? intel 310 40gb, kingston msata 64gb? the second one costs almost the same, but im afraid it has really low random read/write times... cant find any review of this ssd drive tho

 

How do i look up whether my laptop can support mSATA? Am I looking for an mSATA port standard or a pin configuration on the PCI-e board?

Sorry I'm years behind on the developments!! >__<