Thread: Amazing Boot Times!!

I have now tweaked my startup procedure to as much , at this stage, as is possible.

If you have installed Windows 8 in a dual boot, after your existing OS, then, because of it's MBR booting method, Windows 8 will always be the winner. To select Windows 7 from the new boot manager, means the computer will automatically reboot into 7. This adds a penalty of several seconds - on my old test computer 12.
I have, for the purposes of a correct test, reloaded the Windows 7 MBR. My results then were more realistic.

Windows 7 boot 37 seconds. Shutdown 13 Seconds
Windows 8 boot 42 seconds Shutdown 8 seconds.
These figures are taken from the boot manager to a useable computer, including internet connection.

Originally Posted by davehc

I have now tweaked my startup procedure to as much , at this stage, as is possible.

If you have installed Windows 8 in a dual boot, after your existing OS, then, because of it's MBR booting method, Windows 8 will always be the winner. To select Windows 7 from the new boot manager, means the computer will automatically reboot into 7. This adds a penalty of several seconds - on my old test computer 12.
I have, for the purposes of a correct test, reloaded the Windows 7 MBR. My results then were more realistic.

Windows 7 boot 37 seconds. Shutdown 13 Seconds
Windows 8 boot 42 seconds Shutdown 8 seconds.
These figures are taken from the boot manager to a useable computer, including internet connection.

That's why, for now, I'm keeping Windows 8 CP on a HDD of it's own, so that the bootloader isn't a factor. Of course it's going to be a factor, against one OS or the other.

On my notebook, I have no choice in the matter. I need to time it to see which loads the fastest. But then, there I go. Windows 8 CP has the default bootloader, more than likely, it'll boot quicker. But I'll still check.

Cat

Imho, my way is the only true way to test one against the other and, of course, you must have, as far as possible, the same configuration and items loading (Msconfig - startup items, for example). Placing them on separate harddisks, or on different computers, does'nt cut it. If you are certain the HDs are identical, then no big deal, but separate computers will lead into many different parametr problems.

Windows 8 Consumer Preview vs. Windows 7 Start-Up Comparison(VMWare Workstation)

System Specifications:

QuadCore Intel Core i7 Extreme 975, 3466 MHz (26 x 133)
Gigabyte GA-X58A-UD7 v1 (1 PCI, 2 PCI-E x1, 4 PCI-E x16, 6 DDR3 DIMM, Audio, Dual Gigabit LAN, IEEE-1394)
Intel Tylersburg X58, Intel Nehalem, Intel 82801JR ICH10R
24576 MB (DDR3-1333 DDR3 SDRAM)
4 GB DDR3-1333 DDR3 SDRAM (8-8-8-22 @ 609 MHz) (7-7-7-20 @ 533 MHz) (6-6-6-17 @ 457 MHz) 5:1
4 GB DDR3-1333 DDR3 SDRAM (8-8-8-22 @ 609 MHz) (7-7-7-20 @ 533 MHz) (6-6-6-17 @ 457 MHz) 5:1
AMD Radeon HD 6800 Series (1024 MB)
Dedicated storage for test: CRUCIAL_CT128M225 SSD Rev 2030 SATA-II
Host memory: 24576MB
Host OS version: Windows 7 Ultimate, 64-bit 6.1.7601, Service Pack 1


Virtual Machine Settings


VMWare Workstation 8.0.2 build-591240


Memory: 8GB
Processors: 1 processor 4 cores
Hard Drive (SCSI): 60GB (not preallocated, stored in multiple files)
CD/DVD: Not connected on start-up
Network Adapter: Bridged
USB Controller: Present
Sound Card: Present
Display: Auto-detect


Mean Times


Windows 8 CP Start-up Times (in seconds): 38,38,40,37,37 (mean) 38 (average)
Windows 7 w/Service Pack 1 (in seconds): 22,21,21,21,21 (mean) 21.2 (average)

Summary


The start-up test was conducted five times on each system in a virtual machine, running on a 64-bit Windows 7 VMWare Host. During the time of this test, Windows Resource Monitor shows limited to no activity on the host system for disk I/O on all drives, including the system drive, the dedicated program drive, and the drive that hosts VMWare Workstation files. All drives are running CRUCIAL_CT128M225 Rev 2030 SATA-II 128GB SSD.


The Windows 8 CP and Windows 7 64-bit SP1 guest systems ran without any Windows Updates applied. In the case of the Windows 7 64-bit SP1 system, this ran with a slipstream of Windows 7 SP1. Both were clean installs, running VMWare Tools 8.8.2 build-590212. Neither system has had any additional programs installed outside of VMWare Tools. The Windows 7 Consumer Preview installation uses a local account to access the desktop more quickly. Preliminary tests were performed with the Microsoft-connected Live account, which resulted in 40-47 second start times before the virtual machine was recreated.

Result


The test shows an average start-up time of 38 seconds for Windows 8 Consumer Preview and an average start-up time of 21.2 seconds for Windows 7 Service Pack 1. The test was performed five times.


The test shows Windows 8 Consumer Preview loading 16.8 seconds slower than Windows 7 on average under a complete clean install scenario in 64-bit virtualization. This is a 79.245% increase on average. In the best case scenario estimate of a 33 second start time for Windows 8 Consumer Preview, the difference is a 11.8 second slower boot time. This dramatically effects the results, indicating a 55.66% increase as opposed to a 79.245% increase in start-up time. At no time did the test show a decrease in start times on cold boot for Windows 8 Consumer Preview vs. Windows 7 Service Pack 1.


Conditions on the host system did not change in any measurable or substantial way during the test, and neither system was run simultaneously on the same host.


Limitations of Test


It is believed that the VBScript used for this test may be limited: While Windows 7 launched results almost immediately on start-up, Windows 8 Consumer Preview still took around 2-3 seconds after the desktop was displayed to present the results. This could indicate an error rate on the Windows 8 Consumer Preview part of the test, of at most, 5 seconds. It did appear, however, that Windows 8 Consumer Preview was still initiating start-up components of the operating system, even while the desktop was fully rendered, which could have led to legitimately slower start times. In comparison, Windows 7 appeared to be fully loaded and ready to launch any application instantaneously on start-up. It appeared that some processing was taking place while the VBScript results launched (for around 5 seconds) in Windows 8 Consumer Preview. This would give the Windows 8 Consumer Preview test, at most, a 33 second average start-time, when accounting for this difference. The Windows 7 test utilized the same VBScript and encountered no such performance delay.


All hardware and virtual machine settings on both benchmarked systems contained identical settings which did not change at any time during the test. The VMWare Tools for Windows 8 Consumer Preview are the same as the ones installed for Windows 7 Service Pack 1.


These tests utilized identical host-driven hardware and both the host and guests ran under 64-bit architecture. For all intents and purposes, the conditions of the systems were basically identical. VMWare Workstation 8.0.2 build-591240 does not officially support Microsoft Windows 8 Consumer Preview with custom drivers or virtualization enhancements for this system, and runs the operating system as a Windows 7 virtual machine. This is likely to be the closest possible virtualization start-up test for the Windows 8 Consumer Preview at the time the test was conducted on March 13, 2012 with the given software available.


Because the test was performed in a virtual environment, it can not be considered a reliable indicator of real-world results on non-virtualized hardware. unless similar results can be correlated, in terms of percent differential on start time, on similar real-world hardware.


At the very least, Windows 8 Consumer Preview takes 11.8 seconds (55.66%) longer to boot than Windows 7 Service Pack 1, under an entirely clean install, using the latest VMWare Workstation virtualization software. This test is not comparable with non-virtualized hardware or software.

VBScript used (RebootTimer.vbs - 24 lines of code):

Code:

Option Explicit
On Error Resume Next
Dim Wsh, Time1, Time2, Result, PathFile, MsgResult, MsgA, AppName, KeyA, KeyB, TimeDiff
MsgA = "Please close all running applications and click on OK."
KeyA = "HKEY_CURRENT_USER\Software\RestartTime\"
KeyB = "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run\RestartTime"
AppName = "ReBoot-Time"
Set Wsh = CreateObject("WScript.Shell")
PathFile = """" & WScript.ScriptFullName & """"
Result = wsh.RegRead(KeyA & "Times")
if Result = "" then
MsgResult = Msgbox (MsgA, vbOKCancel, AppName)
If MsgResult = vbcancel then WScript.Quit
Wsh.RegWrite KeyA & "Times", left(Time,8), "REG_SZ"
Wsh.RegWrite KeyB, PathFile, "REG_SZ"
Wsh.Run "cmd /c Shutdown -r -t 00", false, 0 
else
Wsh.RegDelete KeyA & "Times"
Wsh.RegDelete KeyA
Wsh.RegDelete KeyB
TimeDiff = DateDiff("s",Result,left(Time,8))
MsgBox "Your computer reboots in " & TimeDiff & " seconds", VbInformation, AppName
end if
wscript.Quit

One area where Windows 8 Consumer Preview does shine is with RAM (memory). This has been mentioned by a few of you and I would concur with this assessment. Thus far, applications seem to be using far less memory, and I believe this may have something to do with the memory deduplication feature that has been introduced into the kernel. I have no proof of this, of course, but I believe that this may be the case. It may also be at least one of the reasons why Windows 8 CP is faster to shut down, but somewhat longer to start. Logically speaking, Windows 8 is loading Metro, which no matter which way you slice it, at least right now, is going to take up more processing time than something like the Start Menu. Not to harp on that issue, though, allocation of memory or the initialization of a newer/better memory handling system for all applications may also be slowing down the startup until it is fully refined. I cannot confirm this or say it is fully accurate, but it is just a hunch. It could also explain why shutdown seems faster - the system is able to offload allocated memory, because ideally, it is already deduplicating it and knows exactly where the assignments are as they relate to file locking (I think). This is all just a guess at this stage, so take my word for a grain of salt on that. It is interesting to see what memory deduplication could do. If it were extended to the file system some day, in real time, it could drastically reduce the amount of file storage being taken up. This technique is already used in the more advanced enterprise backup software suites, but it takes forever to accomplish, and has been known, in many instances, to corrupt a incremental or differential backup (with file data deduplication). I believe that clustering is used to get around the long processing times in large work spaces. Looking up deduplication in general would be a good primer in what I would consider to be, perhaps one of the best technical innovations proposed in this operating system. Unfortunately, I cannot determine where this feature is being used and how well it is functioning from within the OS. I can only see reduced memory usage in comparison with Windows 7.

It could be regarded as a "chicken and egg" situation, Mike. Sinofski's original blog on this, cantered around the lesser requirement for ram, which, even in its latent state, is using appreciable amounts of power. As he said (once again directing us towards the use of Tablets/laptops .. etc ) this would mean less requirement for massive amounts of ram in such equipment, and , axiomatically, less battery consumption.
However, on stationaries today, it is obvious from info on forums, that most users are over endowed with ram on their stationaries so, on those beasts, it is not a factor. I have not tried it to its limits on a laptop, but have seen a couple of articles stating that the battery use is not remarkably improved?
In my own case, my family own several laptops, and have not had a need to put in more ram than comes with the product that includes my game playing grandchildren! On my test stationary, I have 4Gbs. rarely do I exceed 3 when using it, but I do not play games on.

Certainly there are a lot of factors & variables that come into play. To put it in practical, everyday terms, Windows8 is fast, in a nutshell, that's all that really matters. I saw it right away... faster than 7, on a good day... hard to say & does it, matter? It's fast. Boot times are quick but, again, lots of things impact on that, a few milliseconds don't matter & one doesn't boot often, anyway. Still it out of the lab & onto one's desk & in daily working, one only needs to watch the gauges... the computer doesn't (have to) work very hard (or as hard) when running Windows8. Fact... the system resources being tapped are low. Fact... Windows8 will run on not very robust gear. Technically is there a design difference, under-the-covers in Windows8 affecting RAM handling/access, booting & CPU usage? YEP, there sure is. As Dave said, w/ the buckets of RAM in modern machines... In regular use will ppl notice & appreciate all this? Maybe. But, the level of performance is there... noticed & appreciated or not. Features, there are some cool Features... better or worse is a bit subjective & depends on application & environment. Is it good for PDAs & Touch devices? No question. Is it going to help MS compete? Probably. Do ppl need to feel, especially Windows7 users, that they must switch? Nope, not @ all. But, in new machines, does it fit w/ where computing is going? Yep. The novelty of Start will wear off, ppl will learn & adjust...they always have & do. As for all the specs, @ the end of the day... just use it & enjoy it, if you so desire.

Regards,
Drew

I would like to interject and applaud win8 for its speed. I recently built a new desktop with 7200 rpm hd and 3.2gh processor. I installed 8 then ran into an issue only to install seven back over 8. during this process i discovered the answer to the win 8 issue and reinstalled win8 back on my system and windows 8 rocks. very, very, very fast. I liked it so much that i ditched my android phone and got windows phone 7 and its fast as h*** too....way to go microsoft.