Question Performance gains or losses (certain program types?) when disabling windows cache buffer flushing? for both SSDs and HDDs

[GWARslave119]

the title pretty much sums up the question ;P Just real quickly, here's the 3 drives I use:

--------[ Logical Drives ]----------------------------------------------------------------------------------------------

C: (119) Local Disk NTFS 428350 MB 396915 MB 31435 MB 7 % DC79-66C6
D: (119's Slave) Local Disk NTFS 113825 MB 84458 MB 29367 MB 26 % A6F6-9E0D
E: Optical Drive
F: (That's A Huge Bitch) Local Disk NTFS 3725 GB 2536 GB 1189 GB 32 % 80E8-3A5E


--------[ Physical Drives ]---------------------------------------------------------------------------------------------

[ Drive #1 - Samsung SSD 850 EVO 500GB (465 GB) ]

#1 MS Reserved 1 MB 16 MB
#2 Basic Data C: (119) 17 MB 428350 MB
#3 MS Recovery 428368 MB 877 MB

[ Drive #2 - Samsung SSD 850 EVO 120GB (111 GB) ]

#1 MS Recovery 1 MB 529 MB
#2 EFI System 530 MB 100 MB
#3 MS Reserved 630 MB 16 MB
#4 Basic Data D: (119's Slave) 646 MB 113826 MB

[ Drive #3 - ST4000DM000-1F2168 (3726 GB) ]

#1 MS Reserved 1 MB 127 MB
#2 Basic Data F: (That's A Huge Bitch) 129 MB 3815317 MB

And I did a quick read test suite via AIDA64 on 2 drives, 1 before turning off windows write-cache buffer flushing, and 1 after:

By just looking at those, I don't see much of a difference at all between each test, but I also don't know how the program decides to use what Block Size, and since there's different sizes on each test, I don't know if or what the effects the drives would have. My intial guess would be the differences are almost negligible, so no point in turning it off....thoughts?

 

[Palorim12]

the title pretty much sums up the question ;P Just real quickly, here's the 3 drives I use:

--------[ Logical Drives ]----------------------------------------------------------------------------------------------

C: (119) Local Disk NTFS 428350 MB 396915 MB 31435 MB 7 % DC79-66C6
D: (119's Slave) Local Disk NTFS 113825 MB 84458 MB 29367 MB 26 % A6F6-9E0D
E: Optical Drive
F: (That's A Huge Bitch) Local Disk NTFS 3725 GB 2536 GB 1189 GB 32 % 80E8-3A5E


--------[ Physical Drives ]---------------------------------------------------------------------------------------------

[ Drive #1 - Samsung SSD 850 EVO 500GB (465 GB) ]

#1 MS Reserved 1 MB 16 MB
#2 Basic Data C: (119) 17 MB 428350 MB
#3 MS Recovery 428368 MB 877 MB

[ Drive #2 - Samsung SSD 850 EVO 120GB (111 GB) ]

#1 MS Recovery 1 MB 529 MB
#2 EFI System 530 MB 100 MB
#3 MS Reserved 630 MB 16 MB
#4 Basic Data D: (119's Slave) 646 MB 113826 MB

[ Drive #3 - ST4000DM000-1F2168 (3726 GB) ]

#1 MS Reserved 1 MB 127 MB
#2 Basic Data F: (That's A Huge Bitch) 129 MB 3815317 MB

And I did a quick read test suite via AIDA64 on 2 drives, 1 before turning off windows write-cache buffer flushing, and 1 after:

By just looking at those, I don't see much of a difference at all between each test, but I also don't know how the program decides to use what Block Size, and since there's different sizes on each test, I don't know if or what the effects the drives would have. My intial guess would be the differences are almost negligible, so no point in turning it off....thoughts?

I'm not too familiar with this program, but if we are discussing write cache buffer flushing, shouldn't you run a write test? you posted read tests.

 

[GWARslave119]

I'm not too familiar with this program, but if we are discussing write cache buffer flushing, shouldn't you run a write test? you posted read tests.

ok ... well . . . . imagine you walk into a room and see 2 guys, 1 is fervently tapdancing next to his friend who is being gouged in the face by persons unknown during a selfie photo.
.
.
In fact, you don't need to imagine, here:

That's about how I feel right now lol...jesus, the obvious never ceases to go over my head. Ok well, take everything else into account involving the topic question, minus the benchmark photos ;P

 

[GWARslave119]

Oh and this is why I'm not just saying ok lemme do some write test benchmarks right quick ha. Know of any good disk benchmark programs? I have Samsung Magician obviously for the 2 samsung drives, but I was told the benchmarks that it runs is biased towards the drives, something to do with the RAPID technology i think, I do have SiSoft Sandra tho if that's good enough. Or is the way AIDA64 wants to do it the most real-world method of doing write benchmarks?

 

[Palorim12]

ok ... well . . . . imagine you walk into a room and see 2 guys, 1 is fervently tapdancing next to his friend who is being gouged in the face by persons unknown during a selfie photo.
.
.
In fact, you don't need to imagine, here:


That's about how I feel right now lol...jesus, the obvious never ceases to go over my head. Ok well, take everything else into account involving the topic question, minus the benchmark photos ;P

Lol, i've admittedly been out of the SSD game for 2ish years, so i'm brushing up on SSD stuff, but how do the write benches look with it turned on vs off. I texted some of my old coworkers about write cache flushing, so will update when I they respond.

 

[Palorim12]

Oh and this is why I'm not just saying ok lemme do some write test benchmarks right quick ha. Know of any good disk benchmark programs? I have Samsung Magician obviously for the 2 samsung drives, but I was told the benchmarks that it runs is biased towards the drives, something to do with the RAPID technology i think, I do have SiSoft Sandra tho if that's good enough. Or is the way AIDA64 wants to do it the most real-world method of doing write benchmarks?

crystal disk mark, its a classic

 

[GWARslave119]

Sorry for got I posted this ha...I actually didn't really follow through with doing more testing and such, the initial ones I did didn't seem worth it to me to continue on with more testing tbh... I am though avidly fiddling more with o/cing my cpu and ram...more jsut the cpu now, not sure if o/c the ram makes anyt real difference for me..