The Inq: Intel Launching Cheaper 34 nm SSDs in Two Weeks Time

[Denithor]

Originally posted by: halcyon
Correct me if I'm wrong:

1. Gen1 will not get TRIM via firmware upgrade (is this official from Intel?)

For now. They might change that eventually but up front I have a feeling they want everyone who bought into Gen1 to run out and buy a Gen2 drive to stay up with the latest & greatest. Makes sense initially from a purely business sense but long term it doesn't really work to bite the hand that feeds.

2. Gen2 is 2-2.5x faster in 4K random write IOPS than Gen1 (see anandtech)
3. Gen1 due to lack of TRIM can slow down at least 60% (in IOPS / write) after being in use (rec PC Perspective's test on this) compared to a fresh drive
4. Combining the above, in real-life use, Gen 2 can be 5x - 6x as fast as G1 in random write IOPS 4K

Well, they have already released software that can "restore" a drive to the "fresh" state. TRIM simply works to maintain that state - preventing the gradual degradation in performance that occurs without it.

And regarding speeds - I get the feeling we're going to see diminished returns on desktops. Let's say you have a jet that can go 1,000 mph and a second one that can go 5,000 mph. You're not going to see much difference between them if you take 30 mile trips (the equivalent of surfing, watching movies, etc). Now, if you have to fly 3,000 miles you will see a dramatic difference (heavy server loads).

 

[rcpratt]

Originally posted by: jatwell
So, can you buy them anywhere yet??

Haven't seen 'em, I dunno how long it usually takes them to show up...

 

[glugglug]

Originally posted by: halcyon
Correct me if I'm wrong:

1. Gen1 will not get TRIM via firmware upgrade (is this official from Intel?)

I was wondering the same. The press release doesn't specify that 50nm won't get it, it just makes it clear that 34nm will. But maybe Anand got this info from Intel directly or has other sources with the more specific info?

3. Gen1 due to lack of TRIM can slow down at least 60% (in IOPS / write) after being in use (rec PC Perspective's test on this) compared to a fresh drive

I don't think this is still true with the latest firmware. My X25-M still at times gets the 70MB/s writes it did new. I kind of suspect Intel might actually even be throttling it to 70MB/s to keep product differentiation between the M and E drives. I don't think IOPS are throttled though. More interesting the Gen2 M IOPS is higher than Gen1 X25-E.

 

[baddog60]

Originally posted by: jatwell
So, can you buy them anywhere yet??

They are listed on ZipZoomFly.com, but they are out of stock currently. Interestingly, they list both a SATA II and SATA 3.0 version.

 

[taltamir]

SATA I = 1.5 gbps
SATA II = 3.0 gbps
SATA III = 6.0 gbps

Sometimes people write SATAII as SATA 3.0 representing the gbps

 

[jimhsu]

Ordered one on buy.com (no picture, but the model is exact.) http://www.buy.com/retail/product.asp?sku=211487190

They shipped ... we'll see.

 

[fleshconsumed]

Heh, with Windows 7 RTM slated for August 6th I'm tempted to pick one up right now. I know that last firmware update for G1 drives did not require wiping out data, but it's still scary thought if TRIM firmware update will require me to reinstall OS.

 

[glugglug]

Originally posted by: baddog60

Originally posted by: jatwell
So, can you buy them anywhere yet??

They are listed on ZipZoomFly.com, but they are out of stock currently. Interestingly, they list both a SATA II and SATA 3.0 version.

They have the 80GB listed as SATA II and 160GB listed as SATA 3.0

I think its a typo in the 160GB product name.

 

[alcoholbob]

A ~26 decreate in write latency puts the random write performance from around ~50MB/s up to ~60-65MB/s which is damn close to the 70MB/s sequential write performance.

The real question is why does Intel try to differentiate product lines by limiting sequential writes? It makes the X25-M look worse than it needs to against certain competing products.

 

[jimhsu]

Buy.com sold out. That was ... pretty fast.

 

[deputc26]

Originally posted by: jimhsu
Buy.com sold out. That was ... pretty fast.

It appears they are back in stock at buy.com through ANTonline.

 

[Thorny]

A google search shows a few stores that have these in stock in limited quantity. I ordered the last one from Cost Central, should be here Fri. Retailers that listed an ETA had 7/30 as thier estimated available date.

 

[Thorny]

Originally posted by: deputc26

Originally posted by: jimhsu
Buy.com sold out. That was ... pretty fast.

It appears they are back in stock at buy.com through ANTonline.

Look at the price jump though, it was below $230 earlier.

 

[Idontcare]

Originally posted by: Astrallite
The real question is why does Intel try to differentiate product lines by limiting sequential writes? It makes the X25-M look worse than it needs to against certain competing products.

Best as I could figure it has to do with their 5yr absolute minimum lifetime guarantee.

There was a thread here a while back, the OP was regarding SMART attributes on X-25M, in which we dug into the tech docs on the SSD and uncovered some cleverly worded sections regarding the controller monitoring total writes per day and automatically throttles down performance as deemed necessary to ensure the drive's minimum lifespan can never go below 5yrs.

Other than that, I think it is kinda crappy if the 70MB/s write speed is truly artificially capped just to make the SLC drives look all the better for enterprise customers.

 

[tokie]

I would bet it's a combo of not wanting to outperform the X25-E as well as to guarantee the minimum lifespan. Obviously a company with a lot on the line like Intel would not want to have its entrance into a new market hampered by drives crapping out because of excessive use.

 

[deputc26]

Originally posted by: Idontcare

Originally posted by: Astrallite
The real question is why does Intel try to differentiate product lines by limiting sequential writes? It makes the X25-M look worse than it needs to against certain competing products.

Best as I could figure it has to do with their 5yr absolute minimum lifetime guarantee.

There was a thread here a while back, the OP was regarding SMART attributes on X-25M, in which we dug into the tech docs on the SSD and uncovered some cleverly worded sections regarding the controller monitoring total writes per day and automatically throttles down performance as deemed necessary to ensure the drive's minimum lifespan can never go below 5yrs.

Other than that, I think it is kinda crappy if the 70MB/s write speed is truly artificially capped just to make the SLC drives look all the better for enterprise customers.

I honestly don't believe that Intel is artificially capping sequential write speed. I seems to me more a conscious decision to go with higher random writes and lower seq. writes. Though I admit I cannot see a reason why you couldn't have both.

Anand's preview is up http://anandtech.com/cpuchipsets/showdoc.aspx?i=3607

Sequential writes up 10%.

 

[rcpratt]

Ahh, someone else list more at the lower price dangit.

 

[cliftonite]

Ordered the 80 gig at Buy.com

 

[alcoholbob]

Originally posted by: deputc26

Originally posted by: Idontcare

Originally posted by: Astrallite
The real question is why does Intel try to differentiate product lines by limiting sequential writes? It makes the X25-M look worse than it needs to against certain competing products.

Best as I could figure it has to do with their 5yr absolute minimum lifetime guarantee.

There was a thread here a while back, the OP was regarding SMART attributes on X-25M, in which we dug into the tech docs on the SSD and uncovered some cleverly worded sections regarding the controller monitoring total writes per day and automatically throttles down performance as deemed necessary to ensure the drive's minimum lifespan can never go below 5yrs.

Other than that, I think it is kinda crappy if the 70MB/s write speed is truly artificially capped just to make the SLC drives look all the better for enterprise customers.

I honestly don't believe that Intel is artificially capping sequential write speed. I seems to me more a conscious decision to go with higher random writes and lower seq. writes. Though I admit I cannot see a reason why you couldn't have both.

Anand's preview is up http://anandtech.com/cpuchipsets/showdoc.aspx?i=3607

Sequential writes up 10%.

The Indilinix drives have gone through 6 production firmwares. Sequential write performance has gained about 30% from the original firmware while random reads and writes have both increased. I would seriously question Intel's software team if they couldn't effect the sequential write performance of the drive which is at the level of a 5400RPM hard drive currently.

Also the way that X25-Ms have performed on write tests (basically straight-line straddling 70MB/s bar) is extremely suspicious.

 

[taltamir]

Originally posted by: Idontcare

Originally posted by: Astrallite
The real question is why does Intel try to differentiate product lines by limiting sequential writes? It makes the X25-M look worse than it needs to against certain competing products.

Best as I could figure it has to do with their 5yr absolute minimum lifetime guarantee.

There was a thread here a while back, the OP was regarding SMART attributes on X-25M, in which we dug into the tech docs on the SSD and uncovered some cleverly worded sections regarding the controller monitoring total writes per day and automatically throttles down performance as deemed necessary to ensure the drive's minimum lifespan can never go below 5yrs.

Other than that, I think it is kinda crappy if the 70MB/s write speed is truly artificially capped just to make the SLC drives look all the better for enterprise customers.

there is a third part of the equation. It was mentioned that intel has some pretty intense background defragging done to limit the speed loss due to steady state. And that the limit could be related to this, to avoid giving the drive more than it can handle.

 

[taltamir]

wow... 290$ at buy.com it was 22X a few hours ago. now i regret not buying it than...

 

[rcpratt]

Hopefully newegg has it up tomorrow morning.

 

[jimhsu]

Just from gauging the sentiment here and other forums, I thought the 228 price wouldn't last long. I'm sure it'll come down to the 1K price a month from now though...

 

[fleshconsumed]

To be fair 29x price is through antonline, a totally different store. Same kind of situation as with amazon.com where other stores use amazon.com website to sell their products.

 

[taltamir]

i am more curious to see if we could get an indlinx for 1 GB/$