Can my 2006 Laptop use Samsung 860 EVO 500GB SSD?

[kasfruit]

11 hours ago, Mr.Scienceman2000 said:

magician atleast on my experience offered drive health and ability trigger trim to ssd but modern SSD got good garbage collector built in so trim is not neccessary and crystaldiskinfo shows drive health as well

TRIM and garbage collector are two different things, and the alternative tools that supposedly support TRIM commands do not work as intended.

https://arstechnica.com/gadgets/2015/04/ask-ars-my-ssd-does-garbage-collection-so-i-dont-need-trim-right/

I know that XP will run fine with any SSD brand if you don't care about its degradation over time, they are able to re-write thousands of GB's after all.

as a side note, Acer recently got into the SSD market but their SSD don't even have a custom made utility.

https://www.theverge.com/2021/4/27/22405943/acer-ssd-solid-state-drive-nvme-pcie-m2-biwin-storage-ram

 

[we3fan]

Thanks guys, I did some tests on a clean install of Win XP Pro x86 SP2.

Test 1, Time to open .avi (XviD) and DVD .vob video on LG 32LB550B TV with 1 USB port:
HDD WD Elements 500GB Portable USB 3.0: 3 sec
SSD 860 EVO with SATA to USB 3.0 Cable: 3 sec
No difference at all, browsing through folders was also same, I guess the USB speed is limiting factor here.

Test 2, CrystalDiskMark 3.0.4:
HDD 500GB:

SSD 860 EVO:

4.5 times faster Read speed, 4.3 times faster Write speed.

Test 3, Copy Test folder 3.46 GB (mixed files - 1949 files, 30 folders, 64 bytes smallest file, 1.09 GB biggest file) from D: to C:
HDD 500GB: took 3m 18s
SSD 860 EVO: took 46s   (4.3 times faster)

Test 4, Boot Win XP:
HDD 500GB: XP booted in 25s
SSD 860 EVO: XP booted in 18s
I didn't install any programs, I didn't install any drivers, I stopped when I saw the Desktop and didn't wait for any programs to load, that's why the boot time may seem fast.
I think the first 10s of the boot process is same for both and is booting the BIOS (I could be wrong).

Here is another 'SSD on SATA 2' Test on Win 8.1 Pro.

I use SSD for the first time so I have a few questions.
1. Real SSD size is 465 GB. Should I leave 20% (93 GB) free space? If I leave 10% (47 GB) free space will the Read and Write speed be slower?
2. Is it better to leave the free space as unallocated unpartitioned space? Or partition all and make partition for free space but keep it empty and never use it?
3. Is there a program for XP that can TRIM all SSD brands?

Edited May 23, 2021 by we3fan

[Dave-H]

I don't think there's any point in leaving unallocated space on an SSD, or an empty partition, it's physically pretty much like using RAM or a memory stick.
It certainly wouldn't make any difference to the read/write speed I wouldn't have thought, because of the way it works.

There should be some ideas here and here on TRIM programs for XP.
As I have a multi-boot machine I rely on Windows 10 doing the TRIM on my NTFS drives.
It won't work on FAT32 drives but I have a DOS (!) manual TRIM program for that courtesy of the late great Rudolph Loew!

[we3fan]

Thanks Dave :), I appreciate it.

I will try some of these TRIM programs.
If a TRIM program says "TRIM finished successfully", is there a way to verify that the TRIM was indeed successful and properly done?
If we compare 2 TRIM programs for example, can one program trim better than the other? Or the TRIM principle is same for all programs?

I tried one more test, I created new C: partition from XP Install to check the SSD Read and Write speed on badly aligned partition, to my surprise
it was exactly the same, maybe if I continue to use the SSD on badly aligned partition the speed will get worse over time, I don't know. But I
guess it's better if I use the SSD on properly aligned partition.

[Dave-H]

I really don't know about partition alignment, I'm sure others will give a more expert opinion on that, but as far as TRIM is concerned I would have thought a drive is either trimmed or it isn't, I don't think there's any degrees of TRIM, unlike defragmentation.

[we3fan]

I see, OK thanks. I haven't experienced it yet so I don't know how TRIM works exactly, but I guess I will learn in time.

If I use the SSD normally for 3 months and I never use TRIM at all, how will I know if the SSD needs TRIM? Will the Read and Write speed get a lot slower?

[jaclaz]

Let's talk about partition disk alignment (if needed).

It is largely a non-issue that grew much larger than needed/required.

It originated from some very "narrow" cases:

Find some more considerations here:

http://reboot.pro/topic/9897-vistawin7-versus-xp-partitioning-issue/

particularly my post:

http://reboot.pro/index.php?showtopic=9897&p=85960

An actual properly written article about the matter is lost, but you can have it via Wayback Machine:

https://web.archive.org/web/20171111042401/http://www.dcr.net/~w-clayton/Vista/DisappearingPartitions/DisappearingPartitions.htm

and here:

There is no noticeable performance difference (read/write/access speed) on "normal" systems for internal devices, in a top class server even shaving off a few milli-micro-nano-pico seconds from each operation may make sense, but on a normal PC that is largely pointless, you can (and will see) differences on external (USB 2.0) devices such as USB sticks, CF cards and similar, as they are much, much slower and with rotating hard disks, as well, particularly the laptop ones that are slowish, not on internal SATA 3 SSD's that are usually stupidly fast.

BUT on SSD it makes sense to use the "new" (since 2006 or so) MB alignment, as the net effect is slightly (and I mean slightly) less memory cells usage which translates in theory in increased life of the device, though - still - if you can make a device that (with the "wrong" alignment) will last 10 years last (with the "right" alignment) 10 years and two months it is not a life changing result, as after three to five years you will likely anyway change the system or the SSD.

All in all, on internal SSD's it can be "good practice", but you MUST be aware how the disk manager of XP won't like it, a single change in the status (active/not active) of a partition makes likely  that all your logical volumes inside extended will be gone. (they can be recovered, but it is not particularly smart to create the issue), so you either use the "classic" 63 sectors alignment or you shouldn't use the extended partition and logical volumes in it (or - third option - don't use the disk manager or diskpart from XP).

jaclaz

Edited May 27, 2021 by jaclaz

[we3fan]

Thanks jaclaz, great info about partition disk alignment, much appreciated.

Let's see if I understand this correctly, Logical-volumes-will-disappear bug will happen if these 4 conditions are met:
Use Windows XP + Align all Partitions to MB + have Extended Partition with Logical volumes + change active status of partition with XP Disk manager or XP Diskpart.
So if I use Windows XP + I Align all Partitions to MB + I have Extended Partition with Logical volumes, BUT I change active status of partition with a thirt-party program (like MiniTool Partition Wizard), does that mean that the bug will not happen?

My local store only had these 3 SSD Brands: Samsung, A-Data and Kingston. Prices were almost the same (1-2$ difference), so I went for the Samsung (73$), I think SSD prices have dropped a lot and SSDs are very afordable these days.
I wonder if I bought A-Data or Kingston SSD of the same size, would the Read and Write speed be exactly the same on my SATA II Laptop?

[jaclaz]

More or less the issue is the following:
1) the traditional alignment was on "head", which plainly means in 99.9999% of hard disks with a geometry of n/255/63 that the "gap" from the MBR to the beginning of  first (primary) volume was (is) 62 sectors (i.e. 63, the amount of sectors in a "head" minus one, the MBR).

2) the SAME gap happens inside extended partition, between the EMBR (first sector of the extended partition) and the first (logical) volume inside it, AND between the following EMBR ant the relative (logical volumes).

When the alignment "(non-)standard" changed to 1 MB, these gaps became 2047 sectors (i.e. 1048576/512=2048 sectors minus one, the MBR or EMBR).

There is nothing wrong with eoither of the two "conventions".

The bug is in Disk Manager coming with (and likely in the Diskpart version that can be used on) XP.

In order to do what amounts to changing one single byte in the MBR from 0x80 to 0x00  (or viceversa) *somehow* the disk manager "travels" the whole chain of logical volumes inside extended and when the (normally on 1 Mb alignment 2048 (2047+1) value of "sectors before" is encountered the whole EMBR logical volume entry is "wiped" (i.e. overwritten with 00's) but the MBR and thus primary volumes/partitions are not changed, of course this same happens  if other (possibly any) changes in Disk Manager are attempted, not only changing the "active status", I don't think anyone made extensive tests on this.

It is as if there is an implied check on current status of the disk partitioning and when something is not the expected value the entry is wiped (silently).

Different (third party) tools are usually fine (particularly those that - in the same version - run on both XP and Vista/7), as generally they are written to do what they are supposed to do (change a single byte) and not to *somehow* check the consistency of the partitioning scheme of the disk at every run, but of course you cannot be sure-sure until you try the specific tool on the specific system.

As said on the mentioned thread, it is not particularly difficult to find and "undelete" the logical volumes as - luckily - only the first entry in the EMBR is wiped, whilst the second (address of "next" EMBR) is left unchanged, but still it is not exactly "trivial".

About speed, generically speaking when we are talking of storage devices they belong to a "bus", where controllers (and relative drivers and protocols) are involved.

Notwithstanding whatever you read at the time when SATA (SATA I) came out, there was not any particular advantage in speed over good ol' ATA (ex IDE) disks (at the time already at the fastest incarnation of the bus at , i.e. theoretical 133 MB/s vs the - as well theoretical - 150 MB/s of SATA I) because the actual devices (rotating hard disks), both the 7200 rpm and the more economical (and largely used in laptops due to lower power requirements) 5400 rpm were slower than that. (to be fair there was a tiny advantage because of NCQ, Native Command Queing that was available on some SATA disks but not on ATA/IDE ones).

In other words, the bottle neck was the hard disk.

Then faster hard disks came out and the bottleneck became the bus or controller or protocol, so motherboards started getting SATA II.

SATA II (theoretically 300 MB/s) is usually enough to deal with *any* rotating hard disks, the bottle neck is again the mass storage device.

Then came SSD's (that in their SATA version largely outperform SATA II speed) and motherboards started getting SATA III.

SATA III (theoretically 600 MB/s) is enough to deal with *any* (SATA) SSD.

Still SSD (in themselves) can be much faster than what SATA III allows (the bottleneck is again the bus or controller or protocol) so new faster buses (for SSD's) were introduced, direct PciE or Nvme.

Putting a "high end" SSD topping the SATA III standard (like 480-500+ MB/s) on a SATA II bus gives no advantage (the resulting speed will be roughly half of what the device is capable of on a SATA III bus),  a "more common", cheaper SSD (with a speed like 350-400-450 MB/s) is already faster than the SATA II bus, i.e. any speed difference in the device speed is cut off by the bus capacity, i.e. leveled down to the bus max speed.

jaclaz

 

Edited May 28, 2021 by jaclaz

[we3fan]

Thanks again jaclaz.

[kasfruit]

On 5/23/2021 at 1:50 PM, Dave-H said:

I don't think there's any point in leaving unallocated space on an SSD
There should be some ideas here and here on TRIM programs for XP.
As I have a multi-boot machine I rely on Windows 10 doing the TRIM on my NTFS drives.
 

in theory Samsung has already reserved some spacing for over provisioning tasks, this is why the 860 EVO is 500GB not 512GB like the PRO version

but it's recommended to leave that 20% (the more the better) as unallocated space.... the SSD will happily use that free space to paste files given he has no access to the Samsung Magician utility.

I don't think any other utility will perform the TRIM task any better than the official Magician tool....

NO. your Windows 10 is unable to know what files have been deleted in the XP partition so you can't do TRIM through another partition.

I have recently bought a brand new 850 PRO 256GB for $40 + shipping and I am looking at some other setup alternatives such as 850 M2 to DDR2/3 slot

[kasfruit]

On 5/27/2021 at 11:51 AM, we3fan said:

If I use the SSD normally for 3 months and I never use TRIM at all, how will I know if the SSD needs TRIM? Will the Read and Write speed get a lot slower?

if the SSD is big enough you are not likely to notice any performance flop but your SSD without TRIM functionality will just downgrade its lifespan.

fortunately crystaldiskinfo utility still works on XP so you can use it to check your SSD total writes with and without TRIM usage.

 

[Dave-H]

22 hours ago, kasfruit said:

NO. your Windows 10 is unable to know what files have been deleted in the XP partition so you can't do TRIM through another partition.

I presume that's because it's a FAT32 partition.
Why I don't know, as Windows 10 seems to play nicely with my FAT32 partitions in every other respect.
I guess MS assumed that nobody would use a FAT32 fixed disk partition on a Windows 10 system!

[j7n]

What would a 1 MB alignment of partition start achieve? Files would still align to much smaller clusters. Is there a program for testing the performance of a series of possible alignments, in increments of a logical sector, to determine by how many sectors the system needs to be shifted?

[jaclaz]

5 hours ago, j7n said:

What would a 1 MB alignment of partition start achieve? Files would still align to much smaller clusters. Is there a program for testing the performance of a series of possible alignments, in increments of a logical sector, to determine by how many sectors the system needs to be shifted?

The 1 MB is only a round, consistent number, multiple of *everything*.

The actual "typical" cluster size is 4 kb, BUT when we are talking of SSD's or flash devices there is a "page size" (that usually is larger, like 16 kB, 32 kB, 64 kB, 128 kB).

As a matter of fact (and this has briefly happened in Vista times with certain sizes of mass storage devices) a 128 kB alignment is as good as the MB one.

Having the start of the partition (more properly of the volume) on a "round multiple" automatically makes (on NTFS BUT NOT on FAT) any file become "aligned" (to 4kB), set apart very small files, the "bulk" of files will be largely (except for the uneven "tail") be "conforming" to the page size also.

On FAT (16/32/64) it is possible to either offset the partition or manipulate the reserved sectors and/or FAT size so that the first file is aligned (and from then onwards all files will be aligned).

Steve6375 did quite a few tests (on USB sticks where the difference is actually noticeable) and made special provisions for partitioning/formatting in his tools, at the moment reboot.pro (where he shared tests results) is down, here it is (hopefully temporarily) via Wayback Machine:

https://web.archive.org/web/20190530042213/http://reboot.pro/topic/16783-rmprepusb-faster-fat32-write-access-on-flash-memory-drives/

but you can probably read something about the matter on his site:

https://rmprepusb.com/

jaclaz

 

Edited June 3, 2021 by jaclaz