Wednesday, August 13, 2014 – 22:19
This PlexTurbo RAM cache would have to save yourself as well as Samsung be located RAPID. This is all great for high scores in some benchmarks, but less in practice. On top of that is even lied about the operation: [QUOTE] [URL = http: //blog.goplextor.com/ p = 4859?] With PlexTurbo you will not have to search issue: because our new software uses RAM as a write- through cache. It Means did the data will be written to the RAM and the SSD at the same time, and your data will be less vulnerable to the power loss.URL] [/QUOTE] If that would be true, why can because the write rates with RAM be higher than without cache?
One sees, for example, AS-SSD significantly better write performance and AS-SSD overwrites the test of seq. writing the same test file repeatedly. In a write-through cache, however there would not improve performance: [QUOTE] [URL = http: //en.wikipedia.org/wiki/Cache_ (computing)] – Write-through: write is done synchronously Both to the cache and to the backing store. [/URL] [/QUOTE]
In such benchmarks are introduced only the contents of the last write operation then actually write on the SSD, which obviously also happens when you look at the results. That does not work, if you really synchronously writes to the SSD, but only if you later performs the write on the SSD: [QUOTE] [URL = http: //en.wikipedia.org/wiki/Cache_ (computing) ] – write-back (or write-behind): INITIALLY, writing is done only to the cache. The write to the backing store is postponed until the cache blocks Containing the data are about to be modified / Replaced by new content. [/URL] [/QUOTE] In practice, one usually waits to restore not until the space is needed in the cache that would be too risky, but to write then possible zaitnah back the data when just a pause. This then corresponds to what is written there in the German Wikipedia, only that this write buffer just probably can include the entire cache: [QUOTE] [URL = http: //de.wikipedia.org/wiki/Cache#Write-Back] Continuous Writing (write-through)
The block to be written is immediately stored in the next higher level memory. So that the consistency is assured. Thus, the processor does not have to wait each time until the block in the next higher memory level (which is of course slower than the cache) is stored, to use a buffer (write buffer). If it is running full, the processor but [/URL] [/QUOTE] have to stop and wait. Whether that still is Write-Through? Actually, not really, right?
Thursday, August 14, 2014 – 8:43
Thursday, August 14, 2014 – 12:58
Well that’s MSRP, will therefore be cheaper to buy).