First Intel G3 SSD Benchmarks

"Once we have got the news that Intel third-generation SSD codenamed "Postville Refresh"Intel X25-M G3 SSD will be postponed its release to February next year.However,we found out that it is sold quietly at Taobao (Just like Ebay but it is in China) with labeling the benchmark of X25-M G3.

Intel new-generation SSD X25-M G3 will adopt 25nm IMFT MLC NAND chip,new controller chip,with capacities of 80/160/300/600GB.As official news says,the SSD is with read and write speed of up to 250MB/s and 170MB/s,supporting AES-128."

Source: http://en.expreview.com/2010/10/28/intel-x25-m-g3-ssd-debuts-at-taobao/11208.html

 

Because the results are hidden I've calculated something for you.

 

Compared with HDTune RA of OCZ Vertex 2, the Intel G3 is faster with small files.

OCZ Vertex 2 100GB SandForce SF-1200 SSD Review - HD Tune v4.01 - Random Access - Legit Reviews

Compared with Intel G2 80GB, read performance is only slightly improved.

 

I wonder if that's an actual production-level G3, or if it's an ES/pre-production version.

 

Here's another one:

 

Well for a 300GB drive these scores are pretty good!
Plus the price are most likely to go down, so thats good. Hopefully SSD will become more mainstream with Intel's G3!
Thanks for the links

 

How did the folks at Taobao get their hands on a G3 if its not released?

 

"Released" just means that the final samples are available to the general public, but before that, every electronics company makes a bunch of samples and sends them to its partners for feedback because they don't want to manufacture 10 million parts and then find a defect. They do this multiple times and some of these parts tend to disappear somewhere in the pipeline and find their way to reviewers and/or shops. This is technically not legal, but with a few exceptions (Apple being the most obvious), most companies don't think it is worth it to go after them.

 

^^^Engineering Samples.

 

Still not super impressed with the speeds, but they're more than acceptable and that's not really the pressing question anyways. It's about price point and reliability now.

 

Read speeds shows minor improvements, but read speeds were never a problem with Intel G2. It was write speeds that lagged behind the competition.

If real world write speeds and power consumption have been improved, this will be a very good drive.

 

Wow, Phil, did you just go Super Saiyan? From specifications released from the new Sandforce drives increasing power consumption (for increased performance?), I hope Intel does not follow suit.

 

I just noticed the orange letters. Is that explained somewhere?

I don't think Intel and Sandforce will actually release notebook drives with higher power consumption. They know how important it is.

 

Colors of User Title:
Red= Admin
Purple = Lead Moderator
Green = Super Moderator / Moderator
Blue = Regular Registered User
Dark Blue = Senior Registered User
Black = NBR reviewer
Silver = Notebook Reseller

No orange.... Im confused.

Back on topic, I agree that higher power consumption drives clearly aren't the future but the specs seem to say otherwise. Maybe thats just peak consumption and the drive can be erratic?

 

If these G3's are able to keep these speeds sustained, in real world use, then they are already better than any other SSD I have tried.

As to the higher peak power: I say go for it! As long as the drive is allowed to be used in such a way as to allow 'rush to idle', then, overall, we will have a power efficient drive in normal usage, but when we really need performance the drives are not artifically power capped and will be able to deliver the performance I expect them to.

I'm really impressed with the super responsive access times for the small files (about twice as fast as the SandForce based V2). I hope that translates into a drive that 'feels' as fast as its supposed to be (...the Inferno does) and hopefully feels even faster than the Inferno (while keeping the performance at a constant level, with no DuraClass gimmickry present).

At 600GB capacity, I will switch all my mobile workstations over in a blink. And, I'll use the 100GB Inferno as a big and expensive USB drive (w/enclosure).


Thanks Phil for the thread - really looking forward to testing these as soon as I can get a hold of one.

 

That's assuming the idle power consumption is still low. Several users had posted that their Intel G2s didn't give them significant addition to battery life (as compared to the more power efficient Samsung, Indilinx, and Toshiba drives).

 

I haven't noticed significantly different battery life usage with HD's ranging from a 7K200 to a Toshiba 5400 RPM HD, to a Scorpio Blue a Hitachi 7K500, an XT hybrid, Intel G2 and a Patriot Inferno. Less than 5/10 minutes between all those drives in the same VAIO - what I did notice is the heat of the 7K200 though!

Although the SSD promise of better battery life sounds good - I'm not buying them for that aspect alone. I'm looking strictly for better performance/productivity with all other parameters staying ~equal.

Like Phil said above, I don't believe Intel will release new drives with higher power consumption (and, I think mostly at idle will be the same or lower with the smaller chips).

But, knowing Intel, they can (and should) release a drive that has higher maximum power consumption, if the additional power draw is used directly to allow the drive to 'rush to idle'.

Put another way, if I have a two hour battery life (heavy usage) and I get X amount of work done, I won't have a problem if I have a 1 hour battery life and get the same X amount of work done.

I'll still win in the end.

 

Power consumption can have a bigger influence on battery life with smaller, power efficient laptops.

Especially 11-13" ULV notebooks can gain an hour by using a low power SSD.

 

The article lists pricing from $359 to...$1197

That's nothing impressive, hopefully those are preliminary figures. The G2 X25m 160GB can be bought from Best Buy for $420.

To whoever has the coinage to buy a hard drive costing as much as a new gaming laptop, please post pics of your Ferrari(s).

 

RWUK,

It's because they don't have Ferrari's that they'll have the coinage for these drives!

And yes, those prices are preliminary (I'm sure) - let's see the distributors flooded with them and see where the prices drop to.

 

Im guessing around $600 for 300GB and $1200 for 600GB

 

Cloudfire, that is where I see them positioned too - initally.

But they won't move in enough quantities at those prices for Intel/distributiors to stay happy for long.

Even $600 for the 600GB version is too much as a final price - unless the performance is exponentially higher than anything we have currently.

I don't buy tech for the sake of having/bragging about it.

I buy it because it elevates my productivity from where I'm currently at.

So far, no SSD has been able to gain a foothold in my workstations because what they offer is not even close to what is promised (for the money spent).

 

That is absolutely true. I can verify that from when I had the Acer 1830T (i3-430UM) with the ultra-low-voltage CPU. I noticed much less battery life increase switching to an SSD in a laptop using a full-power CPU (i3-370M).

I hope these G3 drives are built to overcome the power-throttling that seems to be inherent with the HM55 chipset. That chipset seems to be everywhere these days. I understand the rationale behind the limitations but I'd really rather have a high-performing SSD and sacrifice a bit of battery life.

I dont' care about productivity OR bragging rights.
I buy it because I like speed. And, speaking of speed, this i3-370M seems to me to be just as peppy and strong as the i7-620M I had in a previous laptop.

 

Real speed increases (not just mere 'snappiness') does equal greater productivity.

 

I'm sure it does. I wasn't contesting that. It's just that I'm not concerned about productivity, per se. I don't use my laptop for any work-related tasks. (I'm retired... work is no longer in my vocabulary as a meaningful word)

Phil: It seems they've created a new (orange) category for you. See if you can find out what it is. Maybe it has something to do with your location.

 

Well the G3 will have twice as much power consumption as the G2 if i remember correctly.

tilleroftheearth:
Do you think it is crazy to pay $600 for 300GB considering the small speed increase?
I think it will take a long time until the price is $1/GB so i don`t really see the 300GB and the 600GB selling so much, atleast not in the beginning.
But the 80/160 will probably because the price on them will be almost like the G2 is today? I really want the 300 GB because i have to remove so much stuff all the time on my 160GB

 

No way I'd pay that much. I chafed at paying $219 for my 128GB Corsair. Also, I wasn't happy with the rather rapid degradation of performance on the Intel G2 drive I had... but maybe I just had a bad drive. The Corsair is holding up well and has given me no problems. It seems that it is a top-seller -- they often run out of them on Newegg.

 

Or maybe it's because I'm the most addicted NBR user

 

Happy halloween.

 

I do hope the prices come down to something more reasonable. I'd like something more than 80GB, as I'm really seeing how it can benefit things and there's really nothing for me to worry about on the write cycles lifespan. Once I'm set up on it, I barely do any writes.

 

Perhaps Phil can go out as an NBR addict for Halloween, covered in hardware manufacturer stickers and CPU & GPU benchmarks printed down the back and front of his shirt

 

I wonder if it was tested on a mobile iCore where there is the IOPS issue? If this is the case with tweaks maybe the 4K results will be even higher.........

 

I don't think we know (for a fact) that the power consumption is doubled, nor how much performance increase we can expect (real world use, not benchies).

As to how much a 300GB SSD drive is worth - I'd say up to the price of the 500GB XT, if not less.

I agree that $1/GB isn't in the near future - but judging by current performance it should be (and less - much cheaper to manufacture than mechanical HD's).

Even a 300GB drive I would find limiting - 600GB will let me use the SSD as I'm currently using a 500GB XT. The extra ~100GB will allow the drive to have as much over-provisioning as it needs (hopefully!).

Right now, I'm at 4GB free space on my Inferno 100GB drive - and I'm not even using it 'seriously'! I need to get a ton of stuff off of it soon.

 

I think I found the answer. Here is another NBR moderator from The Netherlands -- also with orange name.

Why would they do that? Was NBR or the Forum started in The Netherlands?

 

It's off topic, but the forums originated in the states, with admins and the lead moderator in the US. They have a sense of humour and like to do stuff like this on american holidays. You should see christmas and april fools day.

 

Ahh, okay. Orange for Halloween then.
(I wonder if they carve pumpkins in Europe.)

 

So, except the doubled write seq speed and the lower (i hope) prices, do we expect anything new from the new Intels ?

 

Well orange is the national dutch color so who knows.

About the power consumption: I don't believe the initial figures quoted by Anandtech. We'll have to wait until reviews show up.

 

Well, I hope the responsiveness of the G3's are above the SandForce based Inferno as hinted at with the HD Tune access time scores on pg 1 (vs. the HDT scores of the Vertex 2).

Also, I hope that the performance deterioration is 'cured' in these drives when writing huge amounts of data constantly (as my usage seems to do).

Another thing I would like to see is a switch/setting in the Intel SSD Toolbox to allow the drive to operate as a super power efficient drive (when ultimate performance is not an issue) or as a speed demon and battery life be damned setting.

Not only will this allow users to choose their preferred priority for their SSD, but it will also allow Intel to unlock the full potential possible in their 10 channel controller.

Last, but not least - I am looking forward to Intel using a 20/30/40 channel controller depending on the size of the drive offered.

This will finally break away from the expectation born from mechanical HD's that as we get bigger capacity, we get slower drives.

We already have this now with SSD's in small/medium nuances - but I want this at an exaggerated level to partially compensate about paying a premium for simply more capacity.

So, yeah, you could say I'm expecting a little more than an expected speed increase and a price cut.

 

I 2nd that -- but then there's the issue of certain chipsets (HM55/PM55) that seem to scale back the power to the drive. They should modify their chipsets too.

 

@tilleroftheearth : I think that you will have to wait for a couple of years in order to find an SSD to match your needs. For desktop systems there are the FusionIO SSDs that can make you happy (if you have the $$) but for laptops ... none.

 

Well, the Sandforce 2000 chip will give the G3 a beating.
G3:
Sequential Read: 250MB/s
Sequential Write:170MB/s
read IOPS: 50K
write IOPS:40K
Security: AES-128

Sandforce 2000:
Sequential Read: 500MB/s
Sequential Write:500MB/s
read IOPS: 60K
write IOPS:60K
Security: AES-256


Source: http://www.anandtech.com/show/3971/s...and-60k-iops/1

 

ickibar123,

lol... not in a SATA2 system.

Besides, I read Intel's specs as closer to 'minimums' than SandForce spec's which are closer to 'maximums'.

Hope to be proved wrong.

 

Specs don't mean a thing when comparing products that measure them differently. We'll see how they perform in real life situations when they come out, but I doubt they'll be very big differences between the two.

 

wow! if that sandforce 2000 controller is anywhere near that speed it will destroy the intel G3. intel may more fairly measure there speeds... but your looking at a factor of 2, that's an aggressive claim.

considering the lack of significant improvement (in speed) in these G3 results it better not use more power, or I will not be recommend them at all.

I'd be a little more forgiving if they were significantly better preforming but they aren't, at least not according to this article.

 

I think people are forgetting that 'optimistic' doesn't even begin to describe Sandforce's marketing. For a start those transfer rates are theoretical and almost certainly achieved with ATTO which just 0 fills so the data is stupidly compressible. Their IOmeter counts are always optimistic too and lets not forget that basically nobody saw the sequential 285/275 mb/s quoted for SF-1200 even with 0 fill.

SF-1200 was hyped as the second coming of Jesus and whilst it turned out to be good, Sandforce's marketing led consumers to think it was something that it wasn't. Like SF-1200 I'm willing to bet money that the drive is only as fast as the data is compressible. I already think benchmarks are of limited use in sizing up the current SF-1200 drives because all it does is show you the theoretical best and worst transfer rates you can get, with actual speed oscillating wildly in between depending on what type of workload you are subjecting the drive to.

 

Well yeah it's best case scenario but even sequential speeds in itself are a best case scenario. It would be nice if manufacturers would give the PCMark Vantage score but I don't see that happening yet.

I've seen many Sandforce ATTO results equal or very close to the listed specs though. If people don't get very close to it, chances are it's more about the chipset/SATA controller than about the SSD.

Here's an example: http://www.legitreviews.com/article/1299/8/

 

As long as the Intel G3 doesn't perform worse than the G2, is similarly priced, is equally reliable, and has more capacities to choose from, I'll be pretty happy.

 

@ Forge, what if the powerconsumption is now doubled?

 

@Phil, the problem even with PCMark Vantage scores is that the system and methods they test in will still vary too much to mean anything (unless you happen to have the exact same setup).

@yuio, if the performance is the same, the power consumption will be too close to call. If the performance increased substantially, then I don't care if the power consumption goes up accordingly.