Message boards :
Number crunching :
overclocking, seti stability and companies that rape you
Message board moderation
Author | Message |
---|---|
rocketman Send message Joined: 21 Nov 05 Posts: 13 Credit: 56,590 RAC: 0 |
I went past the peak of stability at 2550MHz with my AMD Athlon XP 2600+ Mobile and was getting "client errors" due to a Access violation fault from my firewall's kernel driver(FILTNT.SYS). There have been many prior instances of this kernel driver having these problems but what do you do? According to the people that designed the software, "If your overclocking then that is the problem" end of discussion. The same driver has problems with Hyper Transport and anything that runs over 2000MHz. CPU/FSB speeds are changing fast. If you can't build software that keeps up with these changes then at least have the balls to admit it find some other way to rob people. This is directed at the following thieving bastards: Sun Java, Macromedia, Yahoo, Lavasoft and Agnitumn. They all used to produce good software till the whores sold out. |
J D K Send message Joined: 26 May 04 Posts: 1295 Credit: 311,371 RAC: 0 |
Wrong place to vent your spleen rocketman, complain to them.... And the beat goes on Sonny and Cher BOINC Wiki |
Pappa Send message Joined: 9 Jan 00 Posts: 2562 Credit: 12,301,681 RAC: 0 |
Rocketman Welcome to Seti BOINC While you have issues with the "specific" firewall that you have chosen to install and not being able to keep up with where you would like your computer hardware to be... This is not the place to post other than You have an Issue with Firewall XYZ and Seti BOINC. A Quick look at your two machines show that they may be in need of BIOS Updates as they were not reporting the AMD CPU's correctly. Really if there are issues with the BIOS on the Computer "Many" Software's will function incorrectly including the Basic Operating System. It is also sad that the BIOS is what was installed when you took it out of the box... It is not difficult, but does pose some risk, if you are overclocking then you should also be aware that having the correct BIOS will Make or Break an Overclock. You might also look at Planet AMD for more information. I went past the peak of stability at 2550MHz with my AMD Athlon XP 2600+ Mobile and was getting "client errors" due to a Access violation fault from my firewall's kernel driver(FILTNT.SYS). There have been many prior instances of this kernel driver having these problems but what do you do? According to the people that designed the software, "If your overclocking then that is the problem" end of discussion. The same driver has problems with Hyper Transport and anything that runs over 2000MHz. CPU/FSB speeds are changing fast. If you can't build software that keeps up with these changes then at least have the balls to admit it find some other way to rob people. This is directed at the following thieving bastards: Sun Java, Macromedia, Yahoo, Lavasoft and Agnitumn. They all used to produce good software till the whores sold out. R/ Al Please consider a Donation to the Seti Project. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
I went past the peak of stability at 2550MHz with my AMD Athlon XP 2600+ Mobile and was getting "client errors" due to a Access violation fault from my firewall's kernel driver(FILTNT.SYS). There have been many prior instances of this kernel driver having these problems but what do you do? According to the people that designed the software, "If your overclocking then that is the problem" end of discussion. The same driver has problems with Hyper Transport and anything that runs over 2000MHz. CPU/FSB speeds are changing fast. If you can't build software that keeps up with these changes then at least have the balls to admit it find some other way to rob people. This is directed at the following thieving bastards: Sun Java, Macromedia, Yahoo, Lavasoft and Agnitumn. They all used to produce good software till the whores sold out. When a signal goes from "0" to "1" it does not do so instantly, it takes some time to slew, ring, and then settle. Only then is it safe to decide if the output is zero or one. When you overclock, you are moving from a stable, known point on the waveform closer and closer to a place where 0's are really 0.1, and 1's are more like 0.9. ... and at some point, we're dealing with values dangerously close to 0.5, which can be interpreted as 1 when they're really 0, or 0 when really 1. It seems obvious if you're running the hardware outside the known-good recommended values you will see unexpected (incorrect) results. Regardless of who you decide to call thieving bastards. |
Everette Dobbins Send message Joined: 13 Jan 00 Posts: 291 Credit: 22,594,655 RAC: 0 |
Rocketman > > Rocketman is just venting with a question and you helped him. Hopfully he will take BIOS update advice and be a happy cruncher. |
rocketman Send message Joined: 21 Nov 05 Posts: 13 Credit: 56,590 RAC: 0 |
Geeeez, Don't drink and post. "venting my spleen" good choice of words... I apologize if anyone was offended and there's probably nothing wrong with the products/companies I mentioned. The BIOS for my desktop PC is the latest but will not recognize a "mobile" processor. I have another CPU (3000+) but it has a locked multiplyer and is pretty boring. Interesting point about 1's and 0's possibly being misinterpreted when not given enough time. I was thinking that since only a few, certain programs were malfunctioning it must be poorly written software. I've backed off to 2400MHz and seti/boink is working fine but if I run the driver verifier on filtnt.sys (firewall driver) it will BSOD and flash.ocx (flashplayer) still crashes my Ypager. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
Interesting point about 1's and 0's possibly being misinterpreted when not given enough time. I was thinking that since only a few, certain programs were malfunctioning it must be poorly written software. There is a reason it's called overclocking. Here's a picture: Green is normal clocking. Increase the clock speed, and you move to the left. Yellow is living dangerously. The signal has yet to stabilize, and the black line may "jitter" a bit with temperature. Red is just plain bad. The green line is usually about 5% slower than the fastest reliable speed in testing, because that allows for temperature and voltage variations without moving dangerously toward yellow and red. |
Geek@Play Send message Joined: 31 Jul 01 Posts: 2467 Credit: 86,146,931 RAC: 0 |
I've backed off to 2400MHz and seti/boink is working fine but if I run the driver verifier on filtnt.sys (firewall driver) it will BSOD and flash.ocx (flashplayer) still crashes my Ypager. Do you have the same problems if NOT overclocked? At box-stock speed? Boinc....Boinc....Boinc....Boinc.... |
Jack Gulley Send message Joined: 4 Mar 03 Posts: 423 Credit: 526,566 RAC: 0 |
And over clocking can cause a lot of problems with BIOS code. Often BIOS code has to be written and tweaked for a specific set of system board chip sets. And this usually requires code written for the processor running in a very limited range of speeds, or it will cause the hardware to fail. Sometimes there is just no way around such problems. One thing I have learned with the older AMD Athlon 64 system boards, is not to try over clocking the ones with three memory sockets. Often those do not even allow you to overclock. Memory access will be too unstable. You can usually overclock the boards with just two sockets. But any time you overclock, you have no idea which part of the hardware is going to start failing first. There is not way to tell as each system board and the system board chips on them will all have slightly different tolerances due to the manufacturing processes. When you weight the time, effort and problems recovering from failures, when trying to overclock, with the cost of just buying a faster machine to begin with, unless your time is not worth anything, spending the extra money up front wins out every time. |
Jim Send message Joined: 28 Jan 00 Posts: 614 Credit: 2,031,206 RAC: 0 |
When you weight the time, effort and problems recovering from failures, when trying to overclock, with the cost of just buying a faster machine to begin with, unless your time is not worth anything, spending the extra money up front wins out every time.I hear you but oc'ing is FUN! I really enjoy seeing how far I can take one of my machines. It's the same exact reason why my brother in law spends so much time under the hood of his old Corvette. Tweak, wrench, adjust, swap a part ... it's a hobby. Blowing a cpu (or engine) or having it run rough is just one of the risks you accept as part of the deal. Without love, breath is just a clock ... ticking. Equilibrium |
Alaskan Send message Joined: 18 May 99 Posts: 18 Credit: 53,634,823 RAC: 0 |
Hi, what I have to say about this maybe a repeat to some but here goes anyways. This past weekend, I found at least three issues with my system. The first one was minor but stability in my system was increased immensely, my memory was mismatched, the ras cas etc, was not the same for both pieces of memory. This is very important on a dual channel memory system, so don’t cheap out on the memory. The second issue, the system BIOS was old and required to be updated, not sure if this helped any. Finally the third and biggest issue the system board has a feature called cool and quiet, it’ll slow the CPU fan in an attempt to keep the system noise at a minimum. Switch it off for maximum performance, the cooler the CPU the better your results. Correcting these issues has resulted in a significant improvement in my results, from a 400 average to over 600 credits per day. YMMV but seeing is believing. |
Jack Gulley Send message Joined: 4 Mar 03 Posts: 423 Credit: 526,566 RAC: 0 |
I hear you but oc'ing is FUN! I really enjoy seeing how far I can take one of my machines. It's the same exact reason why my brother in law spends so much time under the hood of his old Corvette. Tweak, wrench, adjust, swap a part ... it's a hobby. Blowing a cpu (or engine) or having it run rough is just one of the risks you accept as part of the deal. Now that you put it that way, I understand and have no problems with someone playing with the hardware to see what it can do. Have done my share of that, but for other reasons. I have only one problem with it in this case. You are playing with real live scientific data. And any glitch caused by your hardware can corrupt the validity of that data and effect the results of others valid work. One of the flaws in the current validation scheme is that if two users of the first three users returning results have errors indicating a noisy WU, then that WU is thrown out as being noisy. Even though one has returned a valid result and the forth in will match it. So thinking there is no harm in returning a bad result from overclocking is not always true. If another user happens to error out on the same WU with you, it can cause a valid WU with potentially interesting data to be rejected as noisy. And there will always be a number of systems that fail and start returning hundreds of these bad results before the owner notices something it wrong. |
Tern Send message Joined: 4 Dec 03 Posts: 1122 Credit: 13,376,822 RAC: 44 |
[quote]One of the flaws in the current validation scheme is that if two users of the first three users returning results have errors indicating a noisy WU, then that WU is thrown out as being noisy. Even though one has returned a valid result and the forth in will match it. So thinking there is no harm in returning a bad result from overclocking is not always true. If another user happens to error out on the same WU with you, it can cause a valid WU with potentially interesting data to be rejected as noisy./quote] This is not correct; to be "valid", two noisy results must be "strongly similar", meaning the two systems would have to error out in the exact same way at the exact same place. If they don't, the validator will wait for a fourth result. If the two _non_ noisy ones then match, all is well. If not, it'll send out a fifth. The other issue is that SETI is not "science" as such, but a search for a signal. If a WU doesn't get analyzed, the only time it would matter is if that was "the signal", which is a chance much smaller than the chance that overclocking would cause a matching error. Some projects don't like overclocked systems; others don't care, because they either have a way to verify real data separately, or the statistics make it irrelevant. |
WR-HW95 Send message Joined: 12 Jan 02 Posts: 10 Credit: 13,513,864 RAC: 0 |
I recomend to test stability of system before actually start to use it after overclocking. http://download.freenet.de/archiv_g/gimps_prime95_3969.html Prime95 custom torture test is good way to see stability....even though it takes time. If it can run over 9 hours, i think it is stable enough for seti. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
When you weight the time, effort and problems recovering from failures, when trying to overclock, with the cost of just buying a faster machine to begin with, unless your time is not worth anything, spending the extra money up front wins out every time.I hear you but oc'ing is FUN! I really enjoy seeing how far I can take one of my machines. It's the same exact reason why my brother in law spends so much time under the hood of his old Corvette. Tweak, wrench, adjust, swap a part ... it's a hobby. Blowing a cpu (or engine) or having it run rough is just one of the risks you accept as part of the deal. Overclocking is fun. When your brother-in-law tweaks on that old 'vette, he is trying to squeeze out the last bit of horsepower, but he's also got to have a car that is reasonably reliable. Fast is cool, but not if you're always towing it back to the garage. Same thing with overclocking a computer: smart overclockers find the fastest speed that "tests clean" and then back down just a little so that they have proper margins: so the signal has time to settle, stop ringing, and be a nice solid reliable 1 or 0 even on hot days or when components have aged. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
The green line is usually about 5% slower than the fastest reliable speed in testing, because that allows for temperature and voltage variations without moving dangerously toward yellow and red. LOL, A great gentle answer to a somewhat silly question/statement. Gee, that probably explains why chip manufactures offer different speed grades in the first place!! Another real world analogy: Race cars generally tend not to be as reliable as the family sedan. Alinator |
spirit_of_the_sixties Send message Joined: 28 Dec 00 Posts: 1 Credit: 339,639 RAC: 0 |
Hi! guys, I think I've really enjoyed reading all the posts about o'clocking and stability etc....and I don't understand a word of it ....I think....I'll just keep on keeping on with my shop bought thingy and hope My new vacuum attachment will suck all the cloggy up dust from the (mysterious) inside of my steam driven word processor thing. I enjoy doing my few 'Cobblestones' a week and it's nice to know there are people out there that are dead brainy and are Really contributing to SetiBoinc.....Onwards and Upwards..to Infinity and beyond!!!! Never mind what others think...if you feel the need, HOWL AT THE MOON!!! |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
The green line is usually about 5% slower than the fastest reliable speed in testing, because that allows for temperature and voltage variations without moving dangerously toward yellow and red. Even so, the race car has to be reliable: if it breaks before the final lap, it won't win even if it's the fastest car out there. |
rocketman Send message Joined: 21 Nov 05 Posts: 13 Credit: 56,590 RAC: 0 |
I've backed off to 2400MHz and seti/boink is working fine but if I run the driver verifier on filtnt.sys (firewall driver) it will BSOD and flash.ocx (flashplayer) still crashes my Ypager. I'm not backing off to box speeds for the 2600 mobile. I did unregister the flash8.dll and am no longer having problems with Ypager closing and have not seen any repercussions as of yet due to this action. My firewall is the new Lavasoft firewall(extremely similiar to Agnitum's Outpost). The only time it BSOD's is when I run Microsoft's driver verifier on FILTNT.SYS and the error is: "FILTNT.SYS was caught trying to corrupt the system". Without Driver Verifier running there are no problems that are detected. RAM is not the problem. I have run extensive tests on the RAM with no errors. What Ned Ludd had to say makes all the sense in the world, And he produced facts to back it up:Not sure how to quote but this is a quote: There is a reason it's called overclocking. Here's a picture: Green is normal clocking. Increase the clock speed, and you move to the left. Yellow is living dangerously. The signal has yet to stabilize, and the black line may "jitter" a bit with temperature. Red is just plain bad. The green line is usually about 5% slower than the fastest reliable speed in testing, because that allows for temperature and voltage variations without moving dangerously toward yellow and red. |
Alinator Send message Joined: 19 Apr 05 Posts: 4178 Credit: 4,647,982 RAC: 0 |
Even so, the race car has to be reliable: if it breaks before the final lap, it won't win even if it's the fastest car out there. You must be a Colin Chapman admirer (founder of Lotus for the unintiated). ;-) His idea of the "perfect" race car was one that went like the devil until it crossed the finish line, and then completely self destructed! :-) Alinator |
©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.