Reprints from my posting to SAN-Tech Mailing List and ...

2011/07/30

[san-tech][03310] "Second-Gen SandForce: Seven 120 GB SSDs Rounded Up", July 25, 2011, Tom's Hardware

Date: Sat, 30 Jul 2011 00:14:47 +0900
--------------------------------------------------
SandForceの 2nd Generation Controller (SF-2500/ 2600) を採用した、
外付け SSDのベンチマークです:

SF-2500/2600
  http://www.sandforce.com/index.php?id=133
  http://www.sandforce.com/userfiles/file/downloads/SandForce_2500-2600ENT_PB_110616.pdf
SandForce
  http://www.sandforce.com/

"Second-Gen SandForce: Seven 120 GB SSDs Rounded Up"
 July 25, 2011, Tom's Hardware
  http://www.tomshardware.com/reviews/sf-2200-sandforce-ssd,2987.html

  "Five SSD vendors sent us seven 120 GB models based on the second-gen
   logic. What makes them different? An extensive benchmark suite tells all."


7機種 (全て 120GB)
  Adata: S511
  Corsair: Force 3
  Mushkin: Chronos Deluxe
  OCZ: Vertex 3
  OCZ: Agility 3
  OCZ: Solid 3
  Patriot: Wildfire
※基本諸元は、"1. A Lot More SandForce" に記載されています。

TABLE OF CONTENTS
  1. A Lot More SandForce
  2. How Can Seven SandForce-Based SSDs Differ?
  3. Test Setup And Firmware Notes
  4. What's Important: Steady-State Performance
  5. Benchmark Results: Storage Bench v1.0 And Real-World Analysis
  6. Benchmark Results: 4 KB Random Performance (Throughput)
  7. Benchmark Results: 4 KB Random Performance (Response Time)
  8. Benchmark Results: 128 KB Sequential Performance
  9. Sequential Performance Versus Transfer Size
 10. PCMark 7: Storage Suite
 11. Final Words

2. How Can Seven SandForce-Based SSDs Differ? に NAND構成比較があります。
Open NAND Flash Interface (ONFI)
  http://onfi.org/

4. What's Important: Steady-State Performance
Steady-State Performance
  "SSD manufacturers prefer that we benchmark drives the way they behave
   fresh out of the box because solid-state drives slow down once you start
   using them. If you give an SSD enough time, though, it will reach
   a steady-state performance level. At that point, its benchmark results
   reflect more consistent long-term use. In general, reads are a little
   faster, writes are slower, and erase cycles happen as slow as you'll
   ever see from the drive."
.....
  "... It's really the only way to examine the true performance of an SSD
   in a way that represents what you'll actually see over time."

  "There are multiple ways to get there, but we're going to use Intel's
   IPEAK (Intel Performance Evaluation and Analysis Kit). This is
   a trace-based benchmark, which means that we're using an I/O recording
   to measure relative performance. Our trace, which we're dubbing Storage
   Bench v1.0, comes from a two-week recording of my own personal machine,
   and it captures the level of I/O that you would see during the first
   two weeks of setting up a computer.

   Installation includes:"
   ※Games, MS Office, Firefox, VMware, Photoshop等

  "The following are stats on the two-week trace of my personal workstation:

    Statistics: Storage Bench v1.0
    Read Operations: 7,408,938
    Write Operations: 3,061,162
    Data Read: 84.27 GB
    Data Written: 142.19 GB
    Max Queue Depth: 452"

※上記 Trace-Basedの Tom's Hardware Storage Bench v1.0以外に

3. Test Setup And Firmware Notes
  Iometer 1.1.0
  ATTO Benchmark
  PCMark 7
※パラメータは上記ページに示してあります。

11. Final Words
Tom's Hardwareとしては、

  "Our 120 GB recommendation goes to the Vertex 3. It costs 20% more than
   the Agility 3 at the same capacity, but its price scales with performance,
   as you can also look forward to getting 22% more performance (at least
   in our Storage Bench v1.0 benchmark)."


0 件のコメント:

コメントを投稿