Message boards :
Number crunching :
Why is the new CC so big?
Message board moderation
Author | Message |
---|---|
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
-Sigh- BOINC is now bloatware? What wonderous new features have been added that make it nearly twice as large as the old CC? Does it now have a built in word processor??? I'm half expecting to find a little game, or photos of the dev team in there like M$ likes to do... :( Don't expect us dial up users to be real quick to upgrade every time a new CC comes out. This project has never been very dial up concious, and it's apparently only going to get worse. |
Neil Walker Send message Joined: 23 May 99 Posts: 288 Credit: 18,101,056 RAC: 0 |
> Don't expect us dial up users to be real quick to upgrade every time a new CC > comes out. This project has never been very dial up concious, and it's > apparently only going to get worse. This is an argument that has raged everywhere since the days of the 110 bps modem. We can't stop progress - nor should we want to. Given the growth of broadband, I don't think the BOINC project's approach is unreasonable - particularly as the whole aim is to move and process massive amounts of data. If you are unable or unwilling to accept the situation as it is, perhaps a smaller DC project might be more to your liking? Be lucky Neil |
Keck_Komputers Send message Joined: 4 Jul 99 Posts: 1575 Credit: 4,152,111 RAC: 1 |
> -Sigh- > > BOINC is now bloatware? What wonderous new features have been added that make > it nearly twice as large as the old CC? Does it now have a built in word > processor??? I'm half expecting to find a little game, or photos of the dev > team in there like M$ likes to do... :( > > Don't expect us dial up users to be real quick to upgrade every time a new CC > comes out. This project has never been very dial up concious, and it's > apparently only going to get worse. > They started using the microsoft MSI installer package that's where the bloat comes in. So there probably are photos of the m$ developers in there. BOINC WIKI BOINCing since 2002/12/8 |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> This is an argument that has raged everywhere since the days of the 110 bps > modem. We can't stop progress - nor should we want to. Yes, this is exactly M$'s way of thinking. People have broadband, HDs now have plenty of storage, why bother to keep things neat and tidy? But is it really progress? > Given the growth of > broadband, I don't think the BOINC project's approach is unreasonable - > particularly as the whole aim is to move and process massive amounts of data. I'm not talking about the size of the WUs, or even the project executables. I'm talking about the BOINC Core Client. I can't fathom why it is so huge. A half hour download for an app that doesn't even do the meat and potatoes work? An app that really only directs traffic? > If you are unable or unwilling to accept the situation as it is, perhaps a > smaller DC project might be more to your liking? Errrmmmm... Yeah... I love this type of answer. Does it really help anyone? |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
> Yes, this is exactly M$'s way of thinking. People have broadband, HDs now have > plenty of storage, why bother to keep things neat and tidy? But is it really > progress? No, it isn't, you're right. > Errrmmmm... Yeah... I love this type of answer. Does it really help anyone? > Yes it helps blowing 'his' horn and make 'him' feel superior for using a real big application ;) @Heffed: I'm with you in this matter. Aloha, Uli |
Neil Walker Send message Joined: 23 May 99 Posts: 288 Credit: 18,101,056 RAC: 0 |
> Yes it helps blowing 'his' horn and make 'him' feel superior for using a > real big application ;) Actually, if you had bothered to look at my computers, you would have seen just how wrong you are. :P I compile my own Linux daemon. Most recent stands at 480998 bytes. ;) Be lucky Neil |
Ulrich Metzner Send message Joined: 3 Jul 02 Posts: 1256 Credit: 13,565,513 RAC: 13 |
> Actually, if you had bothered to look at my computers, you would have seen > just how wrong you are. :P I compile my own Linux daemon. Most recent stands > at 480998 bytes. ;) > Wow, cool, then you were even able to download the real big source code package ;) Aloha, Uli |
Neil Walker Send message Joined: 23 May 99 Posts: 288 Credit: 18,101,056 RAC: 0 |
> Wow, cool, then you were even able to download the real big source code > package ;) Yep. I do it regularly. 4MB at 4Mb/s doesn't take long. :) That's because I threw away my old 1200/75 modem and moved with the times. :P Actually, I am not without understanding for people who want to use dial-up. In fact, I still run a quaint old-fashioned dial-up Fidonet mail system for some of those people and it costs me quite a lot of money to do so. It's horses for courses. I don't feel that BOINC is, or ever has been, the most appropriate project for dial-up users. I see no point in complaining about it, it's a fact of life - the world moves on. Be lucky Neil |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13745 Credit: 208,696,464 RAC: 304 |
> BOINC is now bloatware? What wonderous new features have been added that make it nearly twice as large as the old CC? Keep in mind that the present downloads are still full of debugging code. Once the final versions come out then you'll see a significant drop in the size of the package, as well as a significant drop in the number of new packages being released. Grant Darwin NT |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
> Keep in mind that the present downloads are still full of debugging code. > Once the final versions come out then you'll see a significant drop in the > size of the package, as well as a significant drop in the number of new > packages being released. Even so, many times when you optimize for speed you can get much large output executables. I would trade a lot of space for an increase in speed. Loop unrolling for one can make for much larger code sizes, also putting functions in-line has the same effects. Combine that with residual debug code ... well, haveen't you heard? size doesn't matter ... And if it is that important to you, compile your own .... |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
A large part of the download is the PDB files that can help the developers find the problem when the client goes crash. BOINC WIKI |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> And if it is that important to you, compile your own .... Et tu Paul? Yes, yes, everybody has the required software and knows how to compile their own CC... Not... :( Not to mention as Ulrich brought up, that's an even much larger download. OK, forget I brought it up. Edit: Perhaps I should paraphrase that the only reason I'm concerned about size is the download time on dial up? |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> A large part of the download is the PDB files that can help the developers > find the problem when the client goes crash. I thought those were downloaded when a new project .exe was needed. The previous CC's didn't require these? I remember downloading PDBs for the CC in beta, but they were a separate package. Is JKeck correct that the only reason it's so much bigger is because of the fancy new installer? |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> They started using the microsoft MSI installer package that's where the bloat > comes in. So there probably are photos of the m$ developers in there. HeHe... :) That must be it. |
Neil Walker Send message Joined: 23 May 99 Posts: 288 Credit: 18,101,056 RAC: 0 |
> OK, forget I brought it up. I really was not trying to have a go at you or belittle you in any way. I apologise if it came over that way. I was simply trying to make a constructive suggestion. And I hope that this doesn't sound at all patronising. > Edit: Perhaps I should paraphrase that the only reason I'm concerned > about size is the download time on dial up? I have been involved in dial-up computer comms for many years. My first experience was with a 110 bps acoustic modem connected to a teletype machine so I am well aware of the difficulties. As I mentioned above, I am still actively involved in dial-up comms and have campaigned and still am campaigning for dial-up to remain the default in Fidonet. I remember the time when I had to think very seriously about downloading a 10k file! Right up until last month, I was paying 50ukp per month for an ISDN line I hadn't used in years just so that some (actually one, as it turned out) users could get slightly higher speeds with their analogue modems. I just wonder if dial-up is appropriate for BOINC if you are so concerned (and justifiably) about file size. Be lucky Neil |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
> > A large part of the download is the PDB files that can help the > developers > > find the problem when the client goes crash. > > I thought those were downloaded when a new project .exe was needed. The > previous CC's didn't require these? I remember downloading PDBs for the CC in > beta, but they were a separate package. > The PDB files that I see in my directory are labeled boinc_dll.pdb and boincmgr.pdb each dated 3/3/5. Totaling 4MB. Please remember that there are still bugs that are being chased out of the code, and the PDB files can be used to inject symbols into the error messages so that problems take much less time to debug. (Distributed debugging as well as distributed computing). However, much of the size change is the M$ installer. BOINC WIKI |
Rom Walton (BOINC) Send message Joined: 28 Apr 00 Posts: 579 Credit: 130,733 RAC: 0 |
Without the PDB files the installation package is around 6MB's in size. It's like anything else, cost benefit wise, the extra time it takes to download may help us fix a bug in a day or less, instead of dragging out for months. So much changed under the hood in this release, we didn't want to be stuck without being able to make each crash count and bring us closer to a solution. So far it seems our fears are unfounded, as nobody has reported a crash yet, knock on wood, but it seemed like the right choice. ----- Rom BOINC Development Team, U.C. Berkeley My Blog |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> I just wonder if dial-up is appropriate for BOINC if you are so concerned (and > justifiably) about file size. I'm on disability and can barely afford dial up as it is, so I don't have much choice. I can download the occasional 9+MB file, I was just questioning why it was nearly twice as big when it really doesn't appear to be doing much more than the old one, and was thinking of the recent slew of (almost) daily releases from version 4.14 to 4.19. In a situation like that, instead of upgrading as often as new versions are released, I would tend more towards just turning off BOINC until things were straightened out, which I would hate to do. (I run BOINC 24/7) @ John and Rom, thanks. I guess I'll just have to grit my teeth until the PDBs are gone. (And hopefully new releases won't be rapid) |
Benher Send message Joined: 25 Jul 99 Posts: 517 Credit: 465,152 RAC: 0 |
Anyone out there have an MSI editor? Can anyone tell what (if any) the compression ratios of the various files have? ...they are compresssed...of course....yes? When I compress boinc 4.19 (executable only) it comprises 247K...and thats with my extra CPUID code in it. Haven't done 4.2x yet, so I grabbed the 4.19 versions of 'boinc_gui.pdb', 'boinc_dll.pdb', 'dbghelp.dll' and 'boinc_cli.exe' files, and tried maximum pkzip and 7z compression. Original = 8388k Pkzip = 2495k 7z compress = 2187k Plus M$ MSI overhead =P |
Heffed Send message Joined: 19 Mar 02 Posts: 1856 Credit: 40,736 RAC: 0 |
> Plus M$ MSI overhead =P That would be the photos of whoever wrote the installer, plus maybe a cute game if you press the right button combination while installing. ;) (Ya gotta love M$) But seriously, now you're talking with those compression ratios! :) |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.