Western Digital VelociRaptor Is Launched but what about the Numbers...
by Gary Key on April 21, 2008 12:00 AM EST- Posted in
- Gary's First Looks
Western Digital is finally launching a successor to the Raptor EL150 family of drives that has been around since 2003. This enterprise designed drive series is one that quickly made its home on just about every enthusiast desktop. Capacities started at 36GB, soon grew to 76GB, and then topped out at 150GB for what seemed like an eternity. The new drive appropriately goes by the name, VelociRaptor, and is part of the upcoming VR150 product family. Western Digital is launching this drive at a 300GB capacity point with other capacity points potentially available in the near future.
One of the most interesting aspects of the new drive is its 2.5" form factor. This little wonder packs two 150GB platters featuring 10K-RPM spindle speeds into a 2.5" drive placed in a new 3.5" carrier design dubbed the "IcePAK". This sled features 13 cooling fins that dissipates heat quickly and still allows it to fit in a standard 3.5" drive bay.
We have looked forward to putting this drive through its paces for the past several months since WD first hinted at a Raptor replacement. Our interest grew to a fever pitch last month after being briefed by Western Digital on the VelociRaptor's specifications and knowing a release date had been set for April. We anxiously awaited for test-drives to arrive a couple of weeks ago. As it turns out we received the drive just a few days ago, but were excited at the prospect of a sleepless weekend to see just how well this drive would stack up to the latest competition.
A sleepless weekend is what we got all right, just for all the wrong reasons. We kept noticing the performance of the drive was just not "right". It was certainly faster than its predecessor, but not by the margin, we expected. At times, the benchmark numbers were actually slower than several current 7200RPM drives and this brought out the yellow flag.
Our HDTune and HDTach results kept indicating a massive drop in transfer rates on the outer diameter of the platters, an area where the drive should be the fastest and not the slowest. This pattern continued with our IPEAK and file transfer tests. Eventually, we noticed this speed bump also occurred in several of our application tests where the drive just never seemed to get out of the gates quickly but would eventually catch up with and at times surpass our other drives.
Western Digital informed us the press review samples were engineering drives with early firmware. In fact, Western Digital felt so strongly about this that they included errata in the press kit stating the firmware was about 90% completed. We agree with them. WD confirmed our performance problems and erratic results late yesterday. It turns out the performance drops at the outer diameter are related to the tuning of the servo algorithm. WD had not finalized this tuning when the press sample drives shipped with firmware 03.03VOX.
Western Digital will have this performance drop out fixed before releasing the drives to Alienware next week or retail channels in late May. That said, our “preview” of the drive will be published shortly. We did not want to post a review that offered inconclusive or inconsistent results until we verified problems encountered during testing. As such, we do not plan to “review” the drive until final firmware is available.
Facebook
Twitter
RSS
25 Comments
View All Comments
Gary Key - Monday, April 21, 2008 - link
Actually, the drop offs did make a difference in several application benchmarks, especially file transfers. Now a lot of that will be negated once the drive is loaded with on OS and apps, but our numbers were never consistent. That was the most troubling aspect of testing the drive.What were the true numbers? PCVantage and Nero Recode would have swings around 11~15% as one example. We just felt like it would be prudent to preview the drive now and wait for final firmware before deciding on a conclusion. The preview will be up later today, just with a reduced benchmark suite. We have designed a new benchmark test suite along with testing on AMD/NVIDIA/Intel chipsets now, it was going to be rolled out with this drive, going to have wait a little longer. ;)
The acoustics on this drive are improved a great deal. The idle numbers rank near the WD640 and Samsung 500 and do not stand out. However, the drive can get loud on constant seeks, the noise is muted this time and as such the drive places in the middle of the pack. You will notice it though. We received two new 150GB Raptors last week and I was amazed at the lower acoustics. So much so that we retested the drives for this article.
chizow - Monday, April 21, 2008 - link
Ah I see, I was referring to the HDTune graph you had. I've seen similar drops that did not repeat themselves in test runs with my drives and did not notice any significant drop-off in performance. For example, if you were to average in ~90MB/s for the ms duration the drive dropped down to 8.8MB/s, would it change the overall STR that much? It dropped but that low drop isn't sustained for any significant amount of time.That's good to hear with the improved acoustics. I too was shocked at how much quieter my recent drive back from RMA was compared to my first two Raptor X. I'll be looking forward to updated firmware performance, although at $1/GB it'll have to be really fast given the options out there, as SSD isn't that far off in the $300 price range.
Slash3 - Monday, April 21, 2008 - link
Did your (initial) testing determine that the integrated heatsink was necessary? It seems like a lot of extra flourish for a function that may not be entirely needed. It certainly looks striking, though. I'm anticipating a cringe-worthy price tag to accompany these...Sheninat0r - Monday, April 21, 2008 - link
The MSRP is going to be around $300, so $1/GB.According to some images released by WD, the IcePak cooler reduces drive temperatures by 7C, but in another review it only reduced it by about 4-5C.
Gary Key - Monday, April 21, 2008 - link
It depends on the cooling involved. WD states a 7.1C drop, we measured a 5.1C drop without active cooling on the IcePAK. With active cooling, we had a 8.2C drop with the IcePAK. The temps were still not that bad without it.