本帖最後由 computer1975 於 2011-4-2 07:58 編輯
我重申一次,亦請各位網友留意,SSD是不需要重組的。
Niel 發表於 2011-4-1 21:59


Your OS concept is very poor and do not known well how OS to work!

My standpoint free space fragment hurt the performance of SSD

I remember what myDefrag's autor said. It is the proof of file and space fragment to increase overhead of Windows during I/O operation of SSD.

Memory block fragmentation, filesystem fragmentation, and TRIM
There are 2 kinds of fragmentation that concern SSD disks.

The first kind of fragmentation is memory block fragmentation.
SSD disks are written in pages (generally 4KB in size) but can only be erased in larger groups called blocks (generally 128 pages or 512KB). This causes fragmentation and results in severe performance loss after the disk has been used for a while. Speed can easily drop by 50% or more. The SSD manufacturers have developed a solution called the TRIM instruction, for more information see (http://en.wikipedia.org/wiki/TRIM_%28SSD_command%29)  this Wikipedia article. (kip)…...  the MyDefrag script for Flash memory disks will consolidate free space, and this reduces the problems caused by this kind of fragmentation.

The second kind of fragmentation is filesystem fragmentation. Files can be split into parts that are placed anywhere on the disk, just like on harddisks. Many users think that this kind of fragmentation does not matter for SSD disks, because the disks have a very low latency (no harddisk heads that have to move about). But Windows still has to do more work when a file is fragmented, to gather all the fragments. There is significant overhead inside Windows, nothing to do with the hardware, and it is all the more noticeable because SSD is so fast. MyDefrag deals with this kind of fragmentation.


MyDefrag 有 SSD defrag mode

The MyDefrag script to defragment and optimize Flash memory is specially designed to move as little data as possible. Fragmented files are defragmented (this takes just a single write cycle), unfragmented files are not touched at all. Gaps are filled by moving all the files together (also just a single write cycle), if there are no gaps then MyDefrag will do nothing.

Nevertheless, my advice is to use some discretion and not defragment/optimize flash memory disks every day, but only incidentally, for example once per month.


source: http://www.mydefrag.com/FAQGener ... hSSDMemoryDisk.html

TOP

本帖最後由 Niel 於 2011-4-2 19:23 編輯
Your OS concept is very poor and do not known well how OS to work!

My standpoint free space fragm ...
computer1975 發表於 2011-4-2 07:54 AM


你所有concept用在Hard Drive身上都是對的,不過當你用Hard Drive的concept去了解SSD,所有野都錯曬,我一樣一樣咁解釋你聽:

你的句子:
"My standpoint (is that) free space fragment (fragmentation) hurt(s) the performance of SSD"


我一開始就同意你的講法,不過同時我一開始也說了你的解釋不對。(結論對,解釋不對)可用空間下降引致降速是對的,不過降速不是由File System引致或可用defragment software解決,原因如下:

1. SSD physical layer被controller的FTL abstract左,File system中的logical address已經不是正確的physical address。(不同於HDD)
2. Defragmentation software由於經由File system access physical media,所以Defragmentation SW同樣不能正確運作(因為不能正確地得知physical address,引致不能正確地分析fragmentation)。
3. SSD為了增加速度,是會自行增加fragmentation,目的是使用multichannel的優勢。如果我地假設(留意係假設)有defragmentation software係work,defragmentation只會令SSD降速及增加write wearing,因為defragmentation會抵消了SSD multichannel的效果。幸好的是,現今市場上標榜可以defrag SSD的software都係唔work(原因1及2),所以用左都唔會引致嚴重降速。

現時市場上標榜defrag SSD的software都是defrag file system (above FTL),而不是defrag SSD上的實體data (below FTL)。現時只有TRIM command可以幫助SSD正確地進行block recycle。

我依排好忙,如果你都係唔明的話,我下星期寫一篇詳細分析比大家睇。

TOP

本帖最後由 computer1975 於 2011-4-2 22:44 編輯
我一開始就同意你的講法,不過同時我一開始也說了你的解釋不對。(結論對,解釋不對)可用空間下降引致降速是對的,不過降速不是由File System引致或可用defragment software解決,原因如下:
Niel 發表於 2011-4-2 19:10


不認同 ... ...
1. File System下產生Free Space Fragmentation、File Fragmentation,降速相對上對於JMicro第一代會影響比較太
2. 另外別忘記SSD最麻煩是要erase block之後再能寫東西。Free Space Fragmentation、File Fragmentation長遠來說對SSD有影响,最近出trim技術麻煩少了不少。我所指overhead是Windows處理Free Space Fragmentation、File Fragmentation要額外附出CPU usage,不過當然可以俾錢解決,拿金錢買更快CPU去扺消呢方面I/O latency。
3. 只想純討論SSD沒有再必要爭落去,同時間都有睇下相關filesystem和SSD有乜關係對其他ching可以用,SSD和filesystem兩樣我都有興趣去睇睇。鑽研Filesystem都有10多年啦!

2. Defragmentation software由於經由File system access physical media,所以Defragmentation SW同樣不能正確運作(因為不能正確地得知physical address,引致不能正確地分析fragmentation)。


Diskeeper的HyperFast有針對Free Space Fragmentation作優化的軟件,不過只有benchmark,連一點技術上資料沒有公開過,那要等你或其它ching試下,trial版有一個月免費試用期。本身我只有4G SSD另外也不裝Windows所以無得試。

3. 現今市場上標榜可以defrag SSD的software都係唔work(原因1及2),所以用左都唔會引致嚴重降速。


邊到可以找到資料?
知你不會用MyDefrag的SSD Mode去優化SSD,那就等其他ching去試吧。

1. SSD physical layer被controller的FTL abstract左,File system中的logical address已經不是正確的physical address。(不同於HDD)


認同,本身就是SSD原理

提外話問問ching,你現在用乜CPU、SSD?你係咪Microwork得“將”那位?

TOP

不認同 ... ...
1. File System下產生Free Space Fragmentation、File Fragmentation,降速相對上對於JM ...
computer1975 發表於 2011-4-2 10:15 PM


朋友,我知道你對file system有一定程度的了解,你之前所以說的對於傳統hard drive都是完全正確的。

而且我由第一個回覆就已經說了你的結論是對的 (i.e. 可用空間下降引致降速是對的, 不論HDD or SSD)

不過如果是SSD,成因不是由file system產生,而是由SSD自身的特性引致。而且file fragmentation的產生是對於SSD來說是一個優化手段 (drawback是需要有效的tracing及block recycling,所以才有TRIM command的出現),不過對HDD來說,fragmentation的產生是不想發生的。

Fragmentation的多少及程度控制各廠商都不同,不但沒有標準而且不對外公開 (那些Algorithm好值錢),而且SSD是一個完全abstracted的blackbox(有人比喻為computer inside computer),所以physical layer是完全不受file system管理的。就算是TRIM command,不同的SSD都有不同的處理方法。

所以你說SSD的free space fragmentation及file fragmentation是由OS或FS管理是完全錯誤的。(對於HDD是完全正確)

Diskeeper等Software對第一代的SSD有效,是因為第一代SSD的設計是單純的1-to-1 mapping,那時候Defragmentation software依然可以分析SSD的內部情況,所以效能非常的低。

自Intel x25m第一代之後,主流SSD controller再沒有使用單純的mapping,效能大幅提升之餘,而且已經完全地abstract了。所以現時的SSD不可再被defragmentation software defrag。

時代不同,電腦架構已經完全不同了,舊有的理論已經唔再work。

TOP

我依排好忙,如果你都係唔明的話,我下星期寫一篇詳細分析比大家睇。
Niel 發表於 2011-4-2 19:10


雖然意見不一玫,也沒有敵意,也先行謝過,沒必要去做

I still look for target SSD for 1015PEM notebook.

Kingspec SPF-SF12-S50 (S50) - this one is too pricey
http://www.price.com.hk/product.php?p=124982

Look for Intel G3 - have capacitor but have very short 3000 cycle life

I only use 4G SSD for Linux running on XFS.I tell something to done optimize for my SSD. Flash alignment was done
XFS is made to run in fake RAID mode (512byte x 8 = 4k). Let XFS try to operate to write 4K, 8K, 12K .... each time if possible
XFS is smarter than NTFS to prevent fragmentation because Linux filesystem to prevent fragmentation and delay allocation!

TOP

本帖最後由 Niel 於 2011-4-2 23:24 編輯
雖然意見不一玫,也沒有敵意,也先行謝過,沒必要去做

I still look for target SSD for 1015PEM notebo ...
computer1975 發表於 2011-4-2 11:08 PM


不打緊,我也想大家多了解一點,不過可能要遲一點才有時間寫。

其實3k-5k cycle MLC SSD的壽命是很長的,之前也解釋過:
http://www.hkepc.com/forum/redir ... 24&pid=24691798
所以不用太擔心。請留意計算方法是以日常使用計算 (Intel官方表示WA~1.1),如果你random write多,WA會大幅提升。

看來朋友還對FS及SSD有一點誤解,請你等我下星期空閒些少再給你解釋。

TOP