 |
 |
Why does VirtualPC download faster than OSX??
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Hi,
I've been having an annoying problem. When I download off a certain ftp on Transmit in OSX or any other ftp app I usually get a top speed of 30kbs. But when I download using flashfxp on windows i get 120kbs? Seems a bit strange to me. Does anyone know why this could be.
IT its definately not lying about the speed i've timed it
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Oct 2001
Location: Yokohama, Japan
Status:
Offline
|
|
You sure there wasn't some caching going on? If you did OS X first, and then VPC, you could have been reading from a cache.
Also, I don't know anything about flashfxp, but could it be like SpeedDownload in that it makes multiple simultaneous connections in order to bypass server-side speed limits?
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
hi
thanks for your reply
I'm sure there wasn't any caching as I did download some stuff I hadn't on Transmit. And also I'm sure it doesn't make more than one connection as im not allowed more that 1 on the server  could this be something to do with mtus? could they make this much of a difference?
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
no, because VPC still uses the network services as provided by OS X. In other words, the frames are re-wrapped at the OS X network services level before they ever leave your box.
Did you try the command line FTP client from OS X?
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Senior User
Join Date: Dec 2001
Status:
Offline
|
|
|
|
|
"Think Different. Like The Rest Of Us."
iBook G4/1.2GHz | 1.25GB | 60GB | Mac OS X 10.4.2
Athlon XP 2500+/1.83GHz | 1GB PC3200 | 120GB | Windows XP
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
I'm inclined to believe that, since--like I said--the network traffic out of a virtual machine is routed through the OS X network layer anyway. Barring some client mis-configuration, it's impossible to get faster throughput inside the VM
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Senior User
Join Date: Jun 1999
Location: San Jose, CA
Status:
Offline
|
|
Originally posted by macmike42:
Windows lies (really)
If you are serious, this is easy to tell. Open up teh Activity monitor and switch to teh Network segment. Now fire up VPC, and download something. What does Activity Monitor report seeing for downloads?
|
|
|
| |
|
|
|
 |
|
 |
|
Dedicated MacNNer
Join Date: Nov 2003
Location: Huh?
Status:
Offline
|
|
MenuMeters is your friend. Just have that running and you can tell exactly what speeds you are getting at any given time. Individual apps sometimes fudge their numbers a bit. If there really is a difference though, maybe Transmit is just less efficient or something...dunno. Seems like you shouldn't be losing that much speed because of application inefficiency, and I've never had any problems with Transmit.
|
|
"The captured hunter hunts your mind."
Profanity is the tool of the illiterate.
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
everyone everywhere tells me windows lies. But i am telling you it isnt lying. I've checked activity monitor. My modem even had LED lights which flash more often when downloading faster. It definately downloads faster.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Oct 2001
Location: Yokohama, Japan
Status:
Offline
|
|
None of that matters. Get something like MenuMeters and actually look at what the speed is!
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
If you really want to scientifically analyze the situation, do the following:
Run tcpdump and capture an ftp transfer on OS X.
Run windump and capture an ftp transfer in VPC.
Run Ethereal and analyze the traces. You can even graph packet time, throughput, etc.
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: Edmond, OK USA
Status:
Offline
|
|
A stopwatch works well too. If you need to download a file, you know two very relevant pieces of information: The file size and how long the download took. Record this same data for the VPC Windows ftp client and the Mac OS X ftp command-line client.
Use those numbers to compare, not the values Windows gives.
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Mar 2000
Location: London, UK
Status:
Offline
|
|
It uses evil black magic. Your family is now cursed.
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Originally posted by absmiths:
A stopwatch works well too. If you need to download a file, you know two very relevant pieces of information: The file size and how long the download took. Record this same data for the VPC Windows ftp client and the Mac OS X ftp command-line client.
Use those numbers to compare, not the values Windows gives.
Yup
This is excactly how I know it isn't lying. 900mb in 2 hours in windows. 900mb in 5 hours on OSX.. I really don't want to get into a thing about if its going faster lol I just want to know why it is.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Clinically Insane
Join Date: Nov 1999
Location: 888500128, C3, 2nd soft.
Status:
Offline
|
|
Originally posted by Lakercj:
Yup
This is excactly how I know it isn't lying. 900mb in 2 hours in windows. 900mb in 5 hours on OSX.. I really don't want to get into a thing about if its going faster lol I just want to know why it is.
Chris
Same server?
If you downloaded it in OS X first, and then downloaded the same file in VPC, chances are quite good that it was cached by your ISP.
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
same server, different 900mb file though.. I don't mean to be rude but please belive me it is downloading faster. Im not stupid or new to computers I won't let the evils of windows trick me. I am 120 percent sure its really doing it faster.
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Oct 2001
Location: Yokohama, Japan
Status:
Offline
|
|
For the last freaking time, measure the damn download speed with Activity Monitor or MenuMeters or something!
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Enthusiast
Join Date: May 2002
Status:
Offline
|
|
He's timed it, why the fsck would he need to check the actual DL rate - if it is quicker on a PC, then there must be a reason.
This is not the first time it's been brought up. Maybe Windows is better optimised for broadband downloads, maybe he's downloading off an IIS server and it loves him more on a PC, maybe any number of things.
|
|
I have Mac
|
| |
|
|
|
 |
|
 |
|
Posting Junkie
Join Date: Mar 2001
Location: Salamanca, España
Status:
Offline
|
|
You realize that VPC is an application running in OS X and needs to interact *through* OS X for ALL its I/O operations.
If you are getting 30kBps with Transmit and 120kBps with a Windows ftp app in VPC and your connection to the internet is capable of 120kBps or more then there is only one explanation.
The Windows client is connecting 4 times to the ftp server. The data rate difference even adds up. There is no magic to it, and that is the only explanation.
|
|
I could take Sean Connery in a fight... I could definitely take him.
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
Excuse me, but a stopwatch is far from an accurate analysis of real world network statistics. It doesn't give you anywhere enough information.
Do what I said: tcpdump, windump, and ethereal.
It will tell you exactly what's going on.
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Originally posted by voodoo:
You realize that VPC is an application running in OS X and needs to interact *through* OS X for ALL its I/O operations.
If you are getting 30kBps with Transmit and 120kBps with a Windows ftp app in VPC and your connection to the internet is capable of 120kBps or more then there is only one explanation.
The Windows client is connecting 4 times to the ftp server. The data rate difference even adds up. There is no magic to it, and that is the only explanation.
No, as I said before It can't connect more than one times as im limited to one connection on the server,
I'm about to do the dump thing now.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Oct 2001
Location: Yokohama, Japan
Status:
Offline
|
|
Look, either your perception of the speed is wrong, or you're wrong about the multiple connections thing, or the server is preferentially giving the Windows app higher speeds for no reason. There's no other way around it.
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Originally posted by Kristoff:
Excuse me, but a stopwatch is far from an accurate analysis of real world network statistics. It doesn't give you anywhere enough information.
Do what I said: tcpdump, windump, and ethereal.
It will tell you exactly what's going on.
I tried using tcldump but it just gave me this:
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on en0, link-type EN10MB (Ethernet), capture size 96 bytes
I'm not using an ethernet modem so I don't see how it can work.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Here's my latest attempt to prove it downloads faster though. When I download from the server it gives me my time and speed.
FlashFXP:Transfer complete, 6.5MB in 69 seconds 93.3KB/s
Transmit:Transfer complete, 6.5MB in 177 seconds 36.5KB/s
I did it with Flash first so no caching could occur.
It didnt get up in the 100s on this 1 but still there is a dramatic difference. Remember this data comes from the server not windows or my mac. Now you can say, the activty monitor, windows, my led lights,and the ftp server is lying but I doubt it.
and btw I've now tried the terminal as well.. I get the same results. Also theres no client checking going on or you would see the command on the logs.
Chris
(Last edited by Lakercj; Sep 4, 2004 at 06:18 PM.
)
|
|
|
| |
|
|
|
 |
|
 |
|
Professional Poster
Join Date: Apr 2001
Location: Long Beach, CA
Status:
Offline
|
|
Originally posted by Lakercj:
I tried using tcldump but it just gave me this: I'm not using an ethernet modem so I don't see how it can work.
Chris
If you are using Airport, change en0 to en1.
|

ACSA 10.4/10.3, ACTC 10.3, ACHDS 10.3
|
| |
|
|
|
 |
|
 |
|
Clinically Insane
Join Date: Nov 1999
Location: 888500128, C3, 2nd soft.
Status:
Offline
|
|
Originally posted by GENERAL_SMILEY:
He's timed it, why the fsck would he need to check the actual DL rate - if it is quicker on a PC, then there must be a reason.
You must have missed that it's "quicker" on VIRTUAL PC, running ON THE MAC.
|
|
|
| |
|
|
|
 |
|
 |
|
Clinically Insane
Join Date: Nov 1999
Location: 888500128, C3, 2nd soft.
Status:
Offline
|
|
Originally posted by wataru:
Look, either your perception of the speed is wrong, or you're wrong about the multiple connections thing, or the server is preferentially giving the Windows app higher speeds for no reason. There's no other way around it.
Actually there is:
Transmit could be the culprit?
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Oct 2001
Location: Yokohama, Japan
Status:
Offline
|
|
Yeah, try it with ftp in the Terminal or something.
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
Originally posted by Lakercj:
I tried using tcldump but it just gave me this: I'm not using an ethernet modem so I don't see how it can work.
Chris
RTFM.
man tcpdump
How are you connecting? tell tcpdump which interface to listen on, then set it to capture to a file. Do the same with Windump.
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
And just to dispel any myths people might have regarding OS X being the culprit, here is the output of my ftp of the latest camino build:
ftp> get Camino.dmg.gz
local: Camino.dmg.gz remote: Camino.dmg.gz
227 Entering Passive Mode (64,12,168,21,186,103)
150 Opening BINARY mode data connection for Camino.dmg.gz (7496941 bytes).
100% |************************************************* ********| 7321 KB 479.41 KB/s 00:00 ETA
226 BINARY Transfer complete.
7496941 bytes received in 00:15 (478.53 KB/s)
ftp> by
yeah...that's an average throughput of 478.53 KB/s, without any custom TCP tuning or anything. It's a 10.3.5 box plugged into a NetGear Gigabit Switch uplinked to an ancient BEFSR41 (v1) out through a Linksys cable modem.
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Originally posted by Kristoff:
And just to dispel any myths people might have regarding OS X being the culprit, here is the output of my ftp of the latest camino build:
ftp> get Camino.dmg.gz
local: Camino.dmg.gz remote: Camino.dmg.gz
227 Entering Passive Mode (64,12,168,21,186,103)
150 Opening BINARY mode data connection for Camino.dmg.gz (7496941 bytes).
100% |************************************************* ********| 7321 KB 479.41 KB/s 00:00 ETA
226 BINARY Transfer complete.
7496941 bytes received in 00:15 (478.53 KB/s)
ftp> by
yeah...that's an average throughput of 478.53 KB/s, without any custom TCP tuning or anything. It's a 10.3.5 box plugged into a NetGear Gigabit Switch uplinked to an ancient BEFSR41 (v1) out through a Linksys cable modem.
Thats great and Im vey happy for you but I don't see how thats helping my problem.. I dont like the way everyones having a go at me like I'm against OS X or something or saying Windows is better. I'm not going to try and prove this to you any more It's pointless I've given you a ishload already. I came here for help but instead I just got shotdown,
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: May 2001
Location: Melbourne, Australia
Status:
Offline
|
|
Originally posted by Lakercj:
Thats great and Im vey happy for you but I don't see how thats helping my problem.. I dont like the way everyones having a go at me like I'm against OS X or something or saying Windows is better. I'm not going to try and prove this to you any more It's pointless I've given you a ishload already. I came here for help but instead I just got shotdown,
Chris
Don't stress my man. I believe you, and am intrigued also. I figure it's just the program you are using on each system. Some FTP programs can be slower than others.
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Jan 2003
Location: ~/
Status:
Offline
|
|
Try a different MacOS X FTP program - something other than Transmit, and see what happens. Try it via the FTP capability in Finder (though I wonder if the Finder's FTP is optimized at all).
|
|

|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Jan 2003
Location: NYC
Status:
Offline
|
|
Originally posted by Cadaver:
Try a different MacOS X FTP program - something other than Transmit, and see what happens. Try it via the FTP capability in Finder (though I wonder if the Finder's FTP is optimized at all).
I believe he said he already tried that.
How about ruling out FlashFXP and trying another FTP program on the WIndows side? SmartFTP is good. (sorry if you tried that already)
I found it kind of funny that everyone seemed to get offended that you would suggest something so "blasphemous" as this 
|
"I start fires!"
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2000
Location: in front of the keyboard
Status:
Offline
|
|
Originally posted by Lakercj:
Thats great and Im vey happy for you but I don't see how thats helping my problem.. I came here for help but instead I just got shotdown,
Chris
Dude, I have done nothing but try to help you, so chill.
Read the tcpdump man page, and give it an honest go.
I am telling you right now, everyone can speculate as to what the problem might be, but until you do a packet trace and analyze it with ethereal, you'll never really know what's happening.
|
|
signatures are a waste of bandwidth
especially ones with political tripe in them.
|
| |
|
|
|
 |
|
 |
|
Mac Elite
Join Date: Sep 2001
Location: Chile
Status:
Offline
|
|
use top -d to check network inbound/outbound packages.
|
:: frankenstein / lcd-less TiBook / 1GHz / radeon 9000 64MB / 1GB RAM / w/ext. 250GB fw drive / noname usb bluetooth dongle / d-link usb 2.0 pcmcia card / X.5.8
:: unibody macbook pro / 2.4 Ghz C2D / 6GB RAM / dell 2407wfp - X.6.3
|
| |
|
|
|
 |
|
 |
|
Dedicated MacNNer
Join Date: May 2002
Location: :
:
:noitacoL
Status:
Offline
|
|
I have had this same exact thing happen to me! You are not alone, I believe you.
|
|
|
| |
|
|
|
 |
|
 |
|
Junior Member
Join Date: May 2001
Location: Birmingham,England
Status:
Offline
|
|
|
|
|
Where there's a will
there's a way..
|
| |
|
|
|
 |
|
 |
|
Junior Member
Join Date: Jul 2003
Location: Ontario
Status:
Offline
|
|
Try using command line FTP on the Mac (Yeah I read you tried it, but do it again and time it and all)... then try doing the same on the Windows VPC install, using the windows command line FTP client (Start->Run... ftp ftp.yourserver.com).
Now if there is really a big difference there, that is interesting.
Also, try connecting to a different FTP server.
As for caching, that is silly.... it could possibly be cached on the ISP but that wouldn't be smart of them to do for such a large file... and caching locally? The OS does not cache FTP files  especially TCP/IP connections from other computers (the VPC VM is a separate machine technically). File transfer caching is up to the application.
What about other protocols? Try downloading over HTTP?
Make sure you are not running VPC when you try timing Mac OS X's performance.
|
|
~zig
|
| |
|
|
|
 |
|
 |
|
Professional Poster
Join Date: Nov 2000
Location: Tasmania, Australia
Status:
Offline
|
|
Originally posted by z|gzag:
Make sure you are not running VPC when you try timing Mac OS X's performance.
That's a big one  VPC causes everything else to run slow on the Mac.
VPC also has settings to give it precedence on the processor (using nice, probably) so that when running, the CPU gives more cycles to VPC than to other apps.
|
|
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jun 2003
Location: Canterbury UK
Status:
Offline
|
|
Thanks for all the replys and suggestions guys, I will be sure to spend some time and do them when I get home later on today... stay tuned
|
|
|
| |
|
|
|
 |
|
 |
|
Mac Enthusiast
Join Date: May 2002
Status:
Offline
|
|
Originally posted by Spheric Harlot:
You must have missed that it's "quicker" on VIRTUAL PC, running ON THE MAC.
Rereading my answer I see why it looked like I missed this  , but in fact I was aware that it was VPC - what I meant was people have commented on the fact PCs download quicker before. When running VPC (I haven't run it for a good few years) I believe it reports itself as a PC, certainly on a local network it appears as a new Windows machine, and in server stats it reports it self as IE whatever running on a PC.
If, for instance, a windows client was granted some sort of bandwidth priority when interacting with an IIS server then I think VPC would stand to gain the same sort of advantage.
Similarly, FTP clients I have used with Windows do tend to be better than any of the ones I have used with OSX - I have no trouble believing FlashFXP is a much better client than Transmit (which I have used, and is pretty crap). Maybe it's time to change your OS X FTP client?
|
|
I have Mac
|
| |
|
|
|
 |
|
 |
|
Junior Member
Join Date: Jul 2003
Location: Ontario
Status:
Offline
|
|
Originally posted by GENERAL_SMILEY:
If, for instance, a windows client was granted some sort of bandwidth priority when interacting with an IIS server then I think VPC would stand to gain the same sort of advantage.
This is possible, but highly unlikely... but we all know Mickeysoft and their IIS. Hence, testing on other servers would be a good idea.
|
|
~zig
|
| |
|
|
|
 |
 |
|
 |
|
|
|
|
|

|
|
 |
Forum Rules
|
 |
 |
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is Off
|
|
|
|
|
|
 |
 |
 |
 |
|
 |
|