SEARCH

The How-To Geek Forums Have Migrated to Discourse

How-To Geek Forums / Geek Stuff

Proof - SSD transfer rates are irrelevant for OS performance

(27 posts)
  • Started 1 year ago by whs
  • Latest reply from presence1960
  • Topic Viewed 876 times

whs
whs
Posts: 17584

I think I now have proof that the file transfer rate is pretty irrelevant for the OS performance on a SSD and that all the performance boost comes from the fast access time.

Here is the setup that gave me the clue:

1. I installed Windows 8 in VMware running on my desktop M4 SSD
2. Then I copied the VMware folder to a USB attached Mushkin 60GB SSD
3. I ran Windows 8 off the Mushkin attached via USB3
4. The I ran Windows 8 on the Mushkin via USB2

These are my findings:

a) there seems to be absolutely no performance difference whether I ran on USB3 or USB2 or from the internal M4 SSD

b) what was really surprising was that the boot time (as to Event100 in Event Viewer) was always around 35 sec from the internal M4, but it is 25 sec from the USB3 AND USB2 ports

c) there are a few adjustments that the system and VMware seem to make when you move the VMware folder around. The system made adjustments the first time it detected the moved folder. That was automatic and was only the first time.

VMware seems to make adjustments when you do the first shutdown from the external disk. The Windows 8 shutdown itself was 3 sec (Event 200) but VMware took a couple of minutes to shut everything down - both when I first attached to USB3 and then again on USB2. But that is also only on the first shutdown.

Conclusion: Regardsless whether you attach the SSD on a 5Gb, 6Gb or 480Mb channel, the performance is the same. The transfer speed does not seem to matter.

Installing the OS with VMware on an external SSD is a nice and very flexible option.

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

@whs,

Some R/W Bench Marks would be helpful.

I found this article just doing random search.

"Looking at External Disk Performance using USB 3.0 and eSATA with SSD"

http://www.sqlskills.com/blogs.....-with-ssd/

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

Rick, I was not out for collecting artificial data. This is a real life experience and I was somewhat amazed myself. In the meantime I have booted a few more times both on USB2 and USB3. The boot times vary (don't ask me why) with a spread of 25 sec to 41 sec. But that is still pretty good considering that it runs under VMware.

As far as normal operation is concerned like opening programs, a browser or an OS page, I cannot see any difference to the native performance from my M4 in the desktop. If you measured it, there may be some. But I go by what I see and feel in the real world.

In the meantime, I have also copied the VMware system to a 5400RPM HDD and attached that via USB2. That one is slower, but not too bad. It is very useable with a little delay when calling a page or program first time. But thereafter it is in the RAM cache and goes fast.

I was really testing OS portability using VMware. I run my Windows 8 in VMware and wanted to find a path to carry it to Germany for the summer - rather than starting there from scratch. The acid test is still whether it will run on another system and with a different IP address (because of the activation). I will test the other system variant tomorrow on my laptop and the IP varient on a friend's system. So the performance aspects of the tests are really only a side effect.

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

I understand.

Let us know how it goes on your friend's system.

Posted 1 year ago
Top
 
ispalten
ispalten
Posts: 6259

I sort of have the same feeling WHS... Wife's system has only ONE SATAIII port. Originally came with a 1TB SATAIII drive. Before getting the system I bought a Vertex 4 SATAIII. One of those had to go onto a SATAII port. Originally I put the Vertex on SATAII and then customized the system on the orig. drive copying it over to the SSD sporadically. Booting off of both to make sure all was OK. Of course the SSD screamed and was clearly faster than the mechanical. When satisified all was well, I switched the ports. Although I have no numbers, I really could not tell any apparent difference on the SSD, but the mechanical just seemed to take a little longer to boot.

Booting isn't just about data flowing off or a drive, a lot of other variables come into play, such as device initialization, bus congestion, number of services and processes started, and of course Internet response if connected. It is a chain, and the 'weakest' link is what determines boot speed, especially when all the links are aggregated together. Heck, I've got a program, StartUpCop from PC Mag. that monitors the startup process and will alert me when one app/service it taking too long. I don't get many alerts, but I do get them occasionally. Hardly ever the same one (I think it checks the events).

I also think 'garbage collection' has a lot to do with boot speed, as it might be done prior to system boot (will not appear in event 200). On occasion I see longer delays on W8 before the spinning dots first appear (relative delay compared to other boots).

Irv S.

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

In my case I used the exact same system for all variants. I just copied the VMware folder to the SSD and then booted that from USB3 and USB2.

In the meantime I have also made the test with a 5400RPM spinner on USB2. Now that is really a LOT slower. It runs but it would not be pleasant to operate Windows 8 that way.

I now run Windows 8 from the USB3 attached SSD all the time. It frees up the space on my small 60GB internal M4 and I can run it also on my Toshiba laptop which has a USB3 port, i5 and 6GB of RAM.

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

Feelings are a Good Thing, however they can Never Repeal the laws of Physics and Mathematics.

Until someone comes up with some Firm Bench Marks and Calculations, I'll stick with known mathematical theory.

I'm the product of my profile which I cannot change. (LOL) :) :)

Rick P.

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

Rick, depends of what you are - a Human or an Android. I bet you analyse food by the ingrediants, not by the taste.

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

I read the Specs on everything I eat. (LOL) :) :)

Posted 1 year ago
Top
 
presence1960
presence1960
Posts: 280

You both have good, valid points. I understand where whs is coming from. Initially that is what made me choose AMD over Intel. Benchmark-wise Intel has the decided edge. However in real world with my eyes experience I see no difference in performance. Now I stick with AMD because of many great experiences with their products. As you all well know-that is why there are many flavors of ice cream. To each his/her own. Not to disregard Rick either, for without specs we would be at the mercy of false claims and reckless advertising by hardware manufacturers. Specs are vital because they do tell the truth down to imperceivable increments and help us make decisions about products we want to use.

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

Now you really got Rick going - on AMD versus Intel. Watch this space. LOL

Posted 1 year ago
Top
 
presence1960
presence1960
Posts: 280

That's OK. I like hearing from Rick. If I don't take things personally the man has a lot of good info to share. Besides I did say Intel has the edge benchmark-wise, but with the naked eye I can not perceive a difference. Initially this fact and difference in price made me choose AMD. many years later I still use them, however because of the great experiences I have had with their products which have earned my loyalty. I was not trying debate Intel vs. AMD but rather through examples demonstrate that you both have good, valid points. Just so happened my example of your findings in your experiment was my reason for initially choosing AMD over Intel. OK Rick you can let me have it now :)

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

This tells one all they need to know about the difference between Intel and AMD. :)

Intel Corporation
NASDAQ: INTC - Jan 25 4:00pm ET
20.96 +0.01‎ (0.05%‎)

Advanced Micro Devices, Inc.
NYSE: AMD - Jan 25 4:00pm ET
2.85 +0.06‎ (2.15%‎)

Why AMD's Stock Collapsed And How It Can Recover
http://www.forbes.com/sites/gr.....n-recover/
"While AMD’s chips are competitive at the lower and middle ends of the PC market, it has an insignificant presence in the high margin upper-end of the market."

Posted 1 year ago
Top
 
presence1960
presence1960
Posts: 280

I am not concerned about all that. I only use their products for myself and my customers. As long as their products remain reliable and they put their products on the market I will use them. All I care about is quality and reliability. 99.9% of my customers are not looking for an upper end of the market build anyway. Most people who want that can probably do the job themselves.

Now I am not pushing off what you posted Rick, it is what it is. Their financial performance stinks. Until they succumb and go belly up (then I will be forced to use Intel) or their quality suffers I will continue to use them. BTW in the brief time I have been here you are one of my favorites. You have a lot to offer anyone who is ready to listen to you.

Posted 1 year ago
Top
 
raphoenix
raphoenix
Posts: 14920

OK then we go with Main Stream CPUs and Performance Bench Marks. :)

Just a sample random search.
http://www.tomshardware.com/re.....28-10.html

Posted 1 year ago
Top
 
ispalten
ispalten
Posts: 6259

Don't confuse 'facts' with 'marketing'. Many times the superior solution fails, and it usually is a result of two things, marketing (sales) and adoption. Remember BetaMax vs. VHS? Windows vs. OS/2? Look at the game market?

AMD's problem, trying to be the best low end product I think? How many vendors ship AMD CPU systems? How many vendors actively support and ship Linux configured systems?

Why doesn't Apple have a larger computer share? Simple, closed h/w system, no one else can build them, they are not a 'commodity'.

Irv S.

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

Yeah, Marketing and the Kick backs

Posted 1 year ago
Top
 
ispalten
ispalten
Posts: 6259

Well, sometimes it is the 'maker' who 'shoots themselves in the foot too'. Don't forget Micro Channel Architecture either. No help needed from anyone else, and don't forget the Edsel either!

Irv S.

Posted 1 year ago
Top
 
presence1960
presence1960
Posts: 280

For the things an "average" user does with their machine it makes little difference which brand they have in their box. In today's economic times I lay that right on the line with my customers. I let them know if they want Intel they can have it. But to satisfy their "ego" in lieu of a less expensive CPU that will more than take care of what they need to do with their box is their choice. Most opt for the AMD option for good reason. That's why my customers use me. I am frank, honest and operate with their interests in mind. I find out what they will be using their PC for then recommend a few builds based on those needs. No reason why someone surfing web, email. some photo editing and word processing needs a top of the line Intel CPU. Hardware should fit the customers needs. A little oversell now to make a few more $$ will come back later if they ever realize they didn't actually need that more expensive hardware. Then I lose all future business from that person. Basic rule of sales: ask qualifying questions to ascertain needs. Then present a solution based on those needs.

Posted 1 year ago
Top
 
whs
whs
Posts: 17584

With that approach, I would let you build a new system for me any time. Unfortunately you are not in the geography where I need the new system - the southwest of Germany.

Here in the US I have a recent Dell XPS 8300 with all the bells and whistles. So there is no immediate need for upgrade.

Posted 1 year ago
Top
 



Topic Closed

This topic has been closed to new replies.

Enter Your Email Here to Get Access for Free:

Go check your email!