Here are my results running Performance Test 7 in Windows 7 Home Edition 64b. These figures represent how Parallels scored using my Boot Camp partition for Windows 7 versus Boot Camp being booted natively. Parallels is running fullscreen and isolated from OSX. Macbook Pro 2010 (i7 2.66, 4GB, nvidia330(m)) CPU: 20% slower Memory: 30% slower 2D Graphics: 20% *faster* 3D Graphics: 30% slower Harddrive: 572% slower. Overall that is mighty impressive sans the harddrive performance. I'm not sure what is going on there but that is pretty terrible. Ideas? I have 8GB arriving this week. I will run the tests again when it's installed.
Can you add a virtual hard disk .hdd to your VM and compare performance between the Boot Camp partition and the .hdd? Maybe Boot Camp isn't using caching or something.
Okay WOW! Tested again after importing Boot Camp. MUCH BETTER. I can see from these numbers that Parallels loads most the disk files in to ram giving you a pseudo SSD setup. The hard drive performance is now 88% FASTER than the real install. Also of note is 3D performance is now only 3% slower! VERY GOOD!!! Parallel's 6 is amazing!
I forgot to include the improved CPU number with the imported Boot Camp. It's now only 8% slower than the real install for CPU.
Hello, How did you configure BootCamp and PD6 VM? I mean did you have identical *hardware* configuration of BootCamp and VM? BootCamp Windows use all resources of your Mac by default, i.e. it will use 2CPU + 4GB RAM. But Virual Machine is typically limited to 1CPU and 1GB. So if you have such configuration, then to compare BootCamp and VM correctly you should limit your Windows (in BootCamp) to have just 1CPU and 1GB RAM.
It's also interresting that when you run your BootCamp partition in a VM, you get so slow harddrive performance. Actually harddrive mustn't be slower than native, i.e. it must be on pair. Could you please boot your BootCamp as a VM again and then: 1) re-run Performance Test 2) go to top menu -> Help -> Report a Problem 3) tell me the problem report ID Thank you!
Could you point me to a link explaining how to setup a virtual hard drive .hdd in a VM? Sorry if the question is silly, but I am a beginner in Parallels.
What exactly do you want to do? Convert a BootCamp VM to .hdd or add a new disk to a VM or something else?
Well, I was referring to the performance issue reported above and the suggestion about it made by joevt. I have Parallels installed in a Boot camp partition, and I was wondering if performance would improve if I reinstall it in another way. Is there a way to convert a Parallels based on Bootcamp without reinstalling from scratch? Thanks!
As I mentioned before disk performance in BootCamp VM should be equal to native BootCamp. We need some time to investigate why disk in BootCamp VM could be slower than in native BootCamp. We can't reproduce it so far, but we're working on it. So before convert VM, it would be great to check whether the problem exists on your Mac. To do so, you can download PassMark Performance Test v7 to compare disk performance of native BootCamp vs. BootCamp as a VM. You can download it here: http://www.passmark.com It looks they have a trial version? Then if you will confirm the problem really exists on your Mac also, you can import your BootCamp VM to a .hdd VM: * open VM list (Window -> Virtual Machines List) * do right mouse click (or Command + click) on your BootCamp VM * select "Import" and follow instructions
I did the tests. Please let me know if this seem to be normal or not. R PassMark(TM) PerformanceTest 7.0 Evaluation Version (http://www.passmark.com) Results generated on: Friday, September 17, 2010 Benchmark Results Test Name: This Computer CPU - Integer Math: 263.1 CPU - Floating Point Math: 1091.3 CPU - Find Prime Numbers: 700.4 CPU - Multimedia Instructions: 8.3 CPU - Compression: 3355.0 CPU - Encryption: 12.2 CPU - Physics: 159.3 CPU - String Sorting: 1987.3 Graphics 2D - Solid Vectors: 11.8 Graphics 2D - Transparent Vectors: 10.4 Graphics 2D - Complex Vectors: 110.4 Graphics 2D - Fonts and Text: 130.0 Graphics 2D - Windows Interface: 131.4 Graphics 2D - Image Filters: 366.8 2D Graphics - Image Rendering: 322.0 Graphics 3D - Simple: 59.2 Graphics 3D - Medium: 59.2 Graphics 3D - Complex: 34.6 Memory - Allocate Small Block: 3746.1 Memory - Read Cached: 2271.4 Memory - Read Uncached: 1909.9 Memory - Write: 1362.2 Memory - Large RAM: 946.7 Disk - Sequential Read: 47.4 Disk - Sequential Write: 45.3 Disk - Random Seek + RW: 2.6 CD - Read: 0.8 CPU Mark: 2265.7 2D Graphics Mark: 1068.3 Memory Mark: 796.3 Disk Mark: 344.9 CD Mark: 97.1 3D Graphics Mark: 176.3 PassMark Rating: 916.7 System information: This Computer CPU Manufacturer: GenuineIntel Number of CPU: 1 Cores per CPU: 4 CPU Type: Intel Core i7 M 620 @ 2.67GHz CPU Speed: 2678.1 MHz Cache size: 256KB O/S: Windows XP (32-bit) Total RAM: 2815.5 MB. Available RAM: 2082.0 MB. Video settings: 1920x1123x32 Video driver: DESCRIPTION: Parallels Video Adapter MANUFACTURER: Parallels BIOS: Parallels(R) VGA-Compatible BIOS Version 3.0.2111.89721 DATE: 9-14-2010 Drive Letter: C Total Disk Space: 232.4 GBytes Cluster Size: 4.0 KBytes File system: NTFS
My hardware config was the same for bootcamp and Parallels except for memory. Because when running parallels I need to leave some ram for OSX. Bootcamp sees Windows getting the full 4gb and Parallels was only getting 2gb. However I'm installing 8gb this weekend. I updated my results after importing the Bootcamp partition as a VM hdd file and am getting much better results even with my existing 2gb ram allocation.
ReinaldoG, Did you run Performance Tests in native BootCamp Windows? If so, then you also need to run tests in BootCamp VM under PD6. So we need 2 set of results to be able to compare performance between native Windows and Windows under the PD6
Hi: These are the test results on Bootcamp boot. Please, let me know what you think. Thanks, R PassMark(TM) PerformanceTest 7.0 Evaluation Version (http://www.passmark.com) Results generated on: Friday, September 17, 2010 Benchmark Results Test Name: This Computer CPU - Integer Math: 258.8 CPU - Floating Point Math: 1093.1 CPU - Find Prime Numbers: 708.9 CPU - Multimedia Instructions: 8.4 CPU - Compression: 3409.6 CPU - Encryption: 12.4 CPU - Physics: 182.7 CPU - String Sorting: 2405.6 Graphics 2D - Solid Vectors: 3.8 Graphics 2D - Transparent Vectors: 3.5 Graphics 2D - Complex Vectors: 127.9 Graphics 2D - Fonts and Text: 162.9 Graphics 2D - Windows Interface: 267.8 Graphics 2D - Image Filters: 387.6 2D Graphics - Image Rendering: 327.0 Graphics 3D - Simple: 620.6 Graphics 3D - Medium: 218.0 Graphics 3D - Complex: 52.1 Memory - Allocate Small Block: 3540.2 Memory - Read Cached: 2265.3 Memory - Read Uncached: 2051.9 Memory - Write: 1052.4 Memory - Large RAM: 803.7 Disk - Sequential Read: 50.8 Disk - Sequential Write: 53.4 Disk - Random Seek + RW: 2.8 CD - Read: 0.8 CPU Mark: 2364.3 2D Graphics Mark: 565.1 Memory Mark: 735.3 Disk Mark: 386.9 CD Mark: 94.3 3D Graphics Mark: 554.6 PassMark Rating: 1081.8 System information: This Computer CPU Manufacturer: GenuineIntel Number of CPU: 1 Cores per CPU: 2 CPU Type: Intel Core i7 M 620 @ 2.67GHz CPU Speed: 2660.7 MHz Cache size: 256KB O/S: Windows XP (32-bit) Total RAM: 2230.0 MB. Available RAM: 1714.6 MB. Video settings: 1680x1050x32 Video driver: DESCRIPTION: NVIDIA GeForce GT 330M MANUFACTURER: NVIDIA BIOS: Version 70.16.58.0A.00 DATE: 1-11-2010 Drive Letter: C Total Disk Space: 232.4 GBytes Cluster Size: 4.0 KBytes File system: NTFS
ReinaldoG, you can just compare scores. At first glance Performance Test showed final performance rating of VM is ~10% worse than native. Mostly that happened because of slower 3D in VM. As of disk performance - according to this report VM disk drive performance is on par with native for random access, and 5-10% slower for sequential access. But maybe it's just a test results deviation, or some application on Mac side (Spotlight for instance) also accessed the disk a little at the moment when you run Performance Test in guest. Anyway we'll check what could be wrong with disk performance on BootCamp VM on pixelsmack's Mac, when disk was 500% slower (!?) Thank you all for the reports and independent benchmarking!
Thanks! I also noticed performance is better with Parallels in some specific tests. Another question. Would it have any advantage (performance wise) to import the Parallels from Bootcamp to a VM? R
pixelsmack, just if you're going to run your BootCamp partition as a VM, then could you please re-run Performance Test in it, generate problem report and send me the ID? Thank you!
good! I think BootCamp VM must have the same performance as imported VM, there should be no difference except data location on disk. But with imported VM you will have some other advantages, like normal suspend and snapshots.