| SilverStone SST-EC03 USB 3.0 PCI-E Card |  |
| Reviews - Featured Reviews: Accessories | |
| Written by Marc Fructman | |
| Sunday, 27 November 2011 | |
|
Page 4 of 8
Testing MethodologyPrior to each benchmark the ADATA S511 is zero filled, and the benchmark is run. If the software requires an active partition then this will be performed. A minimal Windows 7 environment is run to ensure repeatability. At least 3 runs are performed for each test. The ADATA N004 Nobility is a working drive with 60% data on it. This SSD is interesting because it ships with both SATA II and USB 3.0 on it.
*Special note: The onboard SATA III of the EVGA SR-2 is powered by the Marvel 9128 Chipset. Please see this article which provides significant detail as to the nature of the problem. For this reason, a Highpoint RocketRAID 2720SGL was used to test baseline speeds of the ADATA S511. Testing and ResultsThe card, bay, and drivers were installed into PCI-E slot #6 on the SR-2 motherboard. A SilverStone TS07 enclosure was used with an ADATA S511 120 GB SSD. An ATTO benchmark was used just to get a baseline idea of transfer speeds. And this is where there were some issues. This card behaved rather oddly depending on the slot that was used. ATTO was showing much lower than expected bandwidth for my first choice slot #6. Everest failed to run at all. PCI Express should not have a problem with data sharing, and the EVGA SR2 has a minimum of x8 lanes for each of the 7 PCI-E slots. Nevertheless, moving from slot #6 to slot #3 or slot #4 made a significant difference in bandwidth speeds, and there was no further problems when using these slots. I suspect this issue was due to an IRQ sharing conflict. For this review we will be using the ADATA S511 120GB SSD and N004 Nobility 64GB SSD / flash drive. The ADATA S511 is reviewed here by Benchmark Reviews. We see that claimed speeds are 550MB/s read and 510MB/s write. The ADATA N004 Nobility is somewhat unique in that it is a flash drive that has both SATA II, and USB 3.0 interfaces built in. According to the manufactures specification it is capable of 230MB/s Seq. Read, and 130MB/s Seq Write on the SATA II interface. For USB 3.0, it claims 180MB/s Seq Read and 150 MB/s Seq Write. Additionally, we will be using the SilverStone TS07 USB enclosure as reviewed here.
|
|
Comments
I may pick up the Silverstone 3.0 for my wife's computer for the price...or wait :)
"The EC03 card from SilverStone installs into a single lane PCI-E gen 2.0 slot..."
How many motherboards built before 2010 even have a "Gen2, 5.0Gbps" x1 PCIe slot? All (or nearly all) only have a Gen1 (2.5Gbps) x1 PCIe slot. Yes, the graphics x16 slot is Gen2 on a vast majority of motherboards, but not the x1 slots.
This also raises serious questions about an inability to use "hubs", "daisy-chaining" or other ways of expanding the number of ports when you run out.
Or... to paraphrase what Bill Gates once said about the PC: "640K should be enough for anybody"... or in this case, "two usb 3.0 ports should be enough for anybody"?
That seems to be a lot of the reason why these boards seem to only carry 2 onboard USB ports, as they are 3rd party solutions.
Intel's new 7X series chipsets do support USB 3.0.
A single PCI Express 2.0 ?lane? can provide 500MB/s in one direction (1000MB/s in both directions).
So, having a 4 port USB 3.0 solution that is plugged in to a single PCI-E slot, is not going to give you full bandwidth for each port. There's just not enough available bandwidth. That is why it might be better for engineers to spec a x4 PCI slot for 4 port USB 3.0
Dunno what AMD's excuse is. Intel said Light Peak, aka Thunderbolt, would obviate the need for it, but their latest X79 chipset doesn't have that, either.