Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel "Emcrest SSDs" Coming....
- Thread starter Wild-Bill
- Start date
- Sort by reaction score
I'm more interested in the G3 Intels. 25nm should bring nice price cuts. I'm actually surprised that Intel rolled out another 34nm SSD, even though their 25nm SSDs should be out during this quarter
Looks good! Of course the Mac Pro is still on SATA II, but still, figure two of these badboys in RAID-0 for the OS, or a PS scratch disk.... Mmmmmm.
Spurce: Fudzilla
Would you even need RAID0 for these? How much "snappier" can Safari be? I see SATA 6.0 SSDs as a means to stop RAIDing SSDs.
Solid state itself is the means to stop striping OS discs for performance.
Striping SSD's will increase the sequential performance (n times the speed of a single drive), but this pretty much linear increase does not apply for random speeds, which is exactly what makes an OS partition on a SSD so fast. The OS doesn't need sequential performance like 500MB/s. Just a quick reference: when I open PS CS5, my Intel drive peaks at 36MB/s sequential speeds, yet it starts within 2 seconds.
raid0 striping can be for money savings. look at two 240gb owc ssd's vs one 480gb in terms of price. 479 and 479 is 958 vs 1579 same space but 600 lower cost.
If your limitation is lack of storage maybe something like this would help. I use an SSD to boot a MBP and have added a mechanical drive using this kit to store VMs.
SSD for apps and HD for video data. That's almost old tech for a MBP configuration.
Johng
Intel must have a lot of left over 32nm parts. 
Looks like I'll be hanging onto my SSD money for a little while longer. My guess is Intel will release there G3 drives right before JMicron releases it's 25nm parts and for a lower price than whatever JMicron sell theres for.
But don't expect much of a price drop from the 25nm parts right away anyway. The SSD companies are saying they have to "recover" the cost of the new SSD controllers first. They should be cheap by Xmas, but thats 11 months away. We will have to pay a similar price per gig as the new Intel 120gb (1.90 per gb in US) for the first few months of the new SSDs.
Looks like I'll be hanging onto my SSD money for a little while longer. My guess is Intel will release there G3 drives right before JMicron releases it's 25nm parts and for a lower price than whatever JMicron sell theres for.But don't expect much of a price drop from the 25nm parts right away anyway. The SSD companies are saying they have to "recover" the cost of the new SSD controllers first. They should be cheap by Xmas, but thats 11 months away. We will have to pay a similar price per gig as the new Intel 120gb (1.90 per gb in US) for the first few months of the new SSDs.
Ah, the wait for the new G3s is driving me nuts. I'm really hoping they'll come out by next month.
Anyway, I at least hope it's "Q1" and not "H1".
I think there is some misleading info at least for our purposes. The ratings are for sata 3. Since the current MacPro does not have sata 3 native, these drives will not do much better if any then the ssd's that are already out. Even with a stat 3 pcie card, the rated numbers will still not happen. I have tested a sata 3 corsair ssd with similar specs to this new Intel and even on the sata 6 pci card it was poorer performance than a sandforce 1200 drive on sata2 slot. If you plug the sata 3 drive on a sata 2 slot, the performance numbers are way worse than the sata 2 sandforce 1200 drives. As you may surmise, I'm not convinced. Sandforce 2000 may change that hopefully. Better yet, maybe next Mac pro will add sata 3 support?
Specs are completely negligible when we talk about two totally different products, with totally different controllers. Just because your Corsair did not achieve the advertised speeds doesn't mean that Intels won't. It's like saying your Ford can't go 200mph because my BMW couldn't...
At least the chipset will support SATA 6Gb/s
You are completely missing the random speeds of the SSD's. Random reads/writes is what makes the SSD so fast for boot and apps, NOT sequential reads.
So even if you put one of the new SSD's on a SATA II controller, you will still see a performance boost in case that the random speeds of that very drive are better than that of your previous drive.
Considering the stated IOPS of that Emcrest drive, it should be considerably slower than the current Postville drives for boot/apps, even though sequential speeds are higher.
However, I don't really trust the stated numbers yet. Those drives are pre-release drives and probably have a non-optimised firmware as of this point.
Unless they post performance specs rated for sata 2, then you simply just don't know if these newer drives are any faster than the current bunch at least as far as on a mac pro is concerned. And I'm not comparing the corsair's marvel chipset with Intel, I'm just saying that a high sata 3 rating doesn't do us much good at the moment. Also the random speeds on the corsair P3 were worse than the random speeds on the vertex 2. Hopefully, that is just on the Corsair's sata 3 drive and won't be a trend.
As per SATA II and SATA III, the difference is the bandwidth of the specification (SATA III doubles that of SATA II). There can be differences in controllers with the same specs (i.e. AMD v. Intel, Marvell, ... located on the system board, not the drive), but it's more on the drive controller than the ICH used in Intel based systems (Intel is the only company that uses their own design that I'm aware of; others are buying parts from 3rd party companies such as SandForce or Samesung... err... Samsung for example
). The specifications as to how the data is transferred are the same (why the controllers can be made backwards compatible in the first place - all they do is throttle the available bandwidth to the slower specification used in the chain). So there will be cases where a SATA II controller will reduce the sequential throughputs as it won't have sufficient bandwidth to handle it (cases where the sequential speed of SATA III based disks can exceed ~2375MB/s, which is the real world limit for SATA II). Burst speeds too, but they're not as important as sustained sequential and random access (even SATA II should be sufficient to handle the Emcrest disks for random access if it does turn out faster than Postville's, which it should once they get the drive's firmware finalized).
BTW, ~70MB/s is all the fastest SSD's currently available can do. So even if this doubles, the SATA II based ICH10 used in the MP can handle it for random access. It's sequential access that the ICH10 will throttle, as the current disks run at the limit of SATA II. It will still work due to backwards compatability in this instance, but it won't run as fast as the drive would be capable of (applicable to most any SATA III SSD that's to release in the near future - the budget models would be the potential exception, such as the Value line from Intel).
I think this is what Transportuer was trying to explain.