Regardless of what option the mods pick as a reason, the recipient is always given the 'obscene' reason. It's a bug as far as I know.

---

Any posts after the first post in this thread may be hidden in this thread. They've been moved here as they were off topic, and I'd made the request for them not to be posted in that thread.

If I can allow someone to imply that I'm lying, and he posts links that prove the point I was making, you should be able to drop this battle of words as well. Let's be respectful of each other's opinions, and agree that we disagree.

I for one have not mentioned you. That is unless you are a member of Boinc Synergy. If you are, then there is no disagreement: I just doubt (for reason) your impartiality.

I worked from the end of the messages back to the beginning. If I didn't move the messages fast enough, then I apologize. I thought it would only take a minute; alas it takes longer than that per each message out of that thread.

If you claim that I'm only for light and fluff - and the opposite is running around attacking each other and each other's teams, then yes, I'm guilty of wanting light and fluff.

I moved the messages using the reason "offtopic". If, in my attempt to move them faster, I managed to choose a different reason, then I'm sorry.

---

Rosetta Moderator: Mod.Tymbrimi
ROSETTA@home FAQ
Moderator Contact

Regardless of what option the mods pick as a reason, the recipient is always given the 'obscene' reason. It's a bug as far as I know.

Actually you are wrong in this. Flamebait is also shown as a reason.
Both subjective terms based on the readers perspective and his/her personal beleifs..

---

Both subjective terms based on the readers perspective and his/her personal beleifs..

That is the definition of "bias" as in anti-XtremeSystems Bias

Any posts after the first post in this thread may be hidden in this thread. They've been moved here as they were off topic, and I'd made the request for them not to be posted in that thread.

If I can allow someone to imply that I'm lying, and he posts links that prove the point I was making, you should be able to drop this battle of words as well. Let's be respectful of each other's opinions, and agree that we disagree.

I for one have not mentioned you. [...]

In message 25971 you quote from my post, and state, "You can moderate and try to rewrite history as you want."

Exactly who are you referring to as "You" that is lying and writing a false description of what happened if it's not the one that wrote the quote in your message.

My message:25967

which refers to this message.


Read Biggles message. Read Tralala's message. The read Hymay's message. Then read my message again, and inform me how my message is different than reality.


Biggles
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2197&nowrap=true#25142

tralala
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2197&nowrap=true#25036

Hymay
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=2197&nowrap=true#25030

I would point out that those machines were perfectly capable of delivering those benches. The std client simply wasn't programmed to use MMX, SSE 1 or SSE2 to compute. Programs are quite literal beasts, If you do not specifically TELL them to use those extensions.. they will not.

Modern cpu's ARE capable of delivering those benchmarks and that power when you program to USE the hardware that is in EVERY modern cpu, from the P2 onward.

Come on.. check your facts.

This is like telling about what my sister in law did when my nephew was born. And then being accused of lying because 3 months prior you described a totally different birth story.

I feel that if you're going to imply that someone's lying that you at least give them the respect to do the research to see what they're talking about.

Alas, I've run out of time to perform the readthrough and requests for cleanup that I'd intended.

---

Rosetta Moderator: Mod.Tymbrimi
ROSETTA@home FAQ
Moderator Contact

This thread, this forum and this project have taken way too much of my time.
It's time to take my efforts elsewhere which I am sure willl please all of you.
You can get back to the people that you want on your project: The do nothings and if that sounds spitefull it is but it is also the truth.
The one's who screamed the loudest for change to get their" cross boinc parity" are also the ones that contributed the least to Rosetta.
You will get the many thousands to assist your project who know nothing of what is really going on and they will be content that they are helping science go forward when the reality is they are helping a project led by a man who doesn't care to take the time to understand the people who donate so much to him.
My eyes have been opened to the type of person he really is.
His own statements say it, I don't have to
Enjoy your world here.
-30-

---

After pulling all 8 of my machines because of the poorly written clients for the Mac, (they only utilize 30% of the CPU making my G5 equal to a G3) I sat back watching to see if anything got better for my eventual return. Reading this and other threads seeing the mods reaction/lack there of, It is quite obvious that,

This project is not fair and will stay that way.
I am not talking about just credits as they may or may not be better now than before.
I am talking about everything including the experience.
The mods step up to defend the accusers after they were attacked by the accused.

This is a project that is, Macs need not apply, we don't need or want you!
This is shown by the new credit system by showing that the client was only capable of using the most basic instructions of the PPC.

Also if you have a lot to contribute we will not defend you but if you defend your self we will stand with the little guy. Why because there are more little guys than big guys.

Be fair to all or lose most.

I pulled all my machines and you (the project) are all the ones that are the losers for it because you lose my contribution, now permanently.

---

@Mod.Tymbrini

Thanks for moving this, it is now easier to follow who responded to whom. I got a message, "Your post was affected by moderation" and the "Off-Topic" reason. Perhaps you should have written, that you just moved the posts and did not delete them, to avoid misunderstandings.

@Movieman

Too bad that you switched back to flame-mode, although we had a profound discussion what went on.

Regardless of what option the mods pick as a reason, the recipient is always given the 'obscene' reason. It's a bug as far as I know.

Ethan: nice try. But from what it has been reported here your always is incorrect.

I have not called Ethan a liar. Lets see how many seconds the post stands.

I aggree with tralala, that the move over in a new thread was appropriate.
But the post by Dave Wilson is imho better suited in the other thread, as it (as I understand it) deals with the flaws (or merits) of the new credit system.

I answer it here nevertheless, but I would appreciate to move both posts over there, as my answer will be OT here:

The scientific application is not suited equally to any CPU architecture. Obviously the newer Macs are not as good at crunching Rosetta as they could be regarding ordinary benchmarks. It boils down to the problem I have discribed in another post in the challenge thread:

there are obviously two different positions for the interpretation of the meaning of credits.
1. A measure for the overall capabilities of a puter (i.e. benchmark) times the donated time.
2. A measure for the work done for the advancement of the science (i.e. delivered decoys)

Both are valid, and in an ideal world both will deliver the same numbers.

The old BOINC method of credit calculation was in favour of #1, but as it was easy to edit the benchmarks without improving the capabilities of the puter, the system was flawed, not to mention it's obvious OS-sensibility.

The second will show how good a certain puter fits in the crunching demands of a certain project. A Ferrari is fine on a racetrack, but off-road he would not succeed. Same goes vice versa for a Hummer. I'm no puter wizz, but that's what I get from the discussion on the different boards: Some puters suit this project better, some another, that's life.

For the different new methods by the projects to calculate the amount of credit, different means are used:
Seti calculates the used Flops within the WU, afaik it would be something more in the direction of #1, only a bit more fiddle-proof.
Folding, Rosetta, CPDN and probably soon enough QMC will give you a certain amount of credit per delivered package of science, it's method #2.

To get the numbers of both methods close enough, it's mandatory that the science application does make use of all the fancy new tricks in the CPU to advance crunching speed, like MMX, SSE(1,2,3) and whatever comes next. I restate, I'm no puter wizz, and I don't know if it's always possible to do so without compromising the results. In calculations that are very sensitive to small rounding differences, it might be crucial not to use the advanced methods, while some other are probably more robust. It's up to the project scientists to decide what suites them best (and what sources of developers to recruit for this task: open or closed source, i.e. volunteers or employers/contractors)

And as I said, optimisation will bring faster results, do not necessary lead to better results. Afaik TeamMacNN is dealing with some of these issues ATM, and perhaps a solution will be possible, but the validity of the results should take the precedence above speed. It's not a game we play, where it doesn't matter whether the shade of green is exactly the same as it was meant to be, but the exact shade of green can essential in the design of a drug.

So the use of MMX, SSE and other stuff might be impossible, but I can't decide this. If it's so, a G5 may really be not faster as his older brother regarding this project, and it comes down to a decision between my mentioned options 1 and 2, and no way of getting them equal in numbers. My opinion: stick to #2, there are enough DC-projects with equal goals and merit who will probably use your puter better, and this is the way to find out how your puter suits the needs of the project, #1 will imho be a waste of CPU-circles, as they can't use the full power in this machine here.

Regardless of what option the mods pick as a reason, the recipient is always given the 'obscene' reason. It's a bug as far as I know.

Ethan: nice try. But from what it has been reported here your always is incorrect.

I have not called Ethan a liar. Lets see how many seconds the post stands.

Hi Jose,

to challenge your prejudice that the mods are biased this post will stand. ;-)
To clarify Moviemans posting history:
In the last 15 days four posts from him were deleted as "flamebait" (which they were) and one as "obscene". The "obscene" one was an empty post and since the reason "obscene" is the default one when deleting a post it was probably not changed since anybody should understand the reason that an empty post was deleted, right?

So here another conspiracy theory of you ends in smoke and you should stop accusing the mods of biased moderation. I checked your last 12 posts and they are all negative and just accusing either the mods or the staff of bias against XS. So please stop it.

---

I am a forum moderator! Am I?

I aggree with tralala, that the move over in a new thread was appropriate.
But the post by Dave Wilson is imho better suited in the other thread, as it (as I understand it) deals with the flaws (or merits) of the new credit system.

Actually my post, in my opinion, was about the apparent lack of respect the project has for me and my Macs because in my opinion, the client was not written with interest in regard to the PPC and it's features. I do not agree that it is acceptable to provide a client that is so flawed that it will make me waste 2 thirds of my processor time and investment thinking I am doing something.

If the client can not produce more that 1 third usage of a Mac PPC processor Don't distribute it.
I have seen other projects that do not have a Mac client.

Those projects don't put out a client that barely works and say here you run this.

So I think with respect this is the correct place for my post.

---

The one's who screamed the loudest for change to get their" cross boinc parity" are also the ones that contributed the least to Rosetta.

As far as I was concerned we were having a perfectly civil and reasonable debate assessing the proposals initially - no screaming involved!

I know you know that cross BOINC parity isn't the reason for the new credit system. I have no particular desire for cross project BOINC parity; I understand the reasons that it is desirable (and can be argued as both necessary and impossible, but that's another topic!), but isn't a concern of mine. Yet you consistently state that the vocal minority consisted of one or two teams that got 'their way' with the new credit system as the results now give similar credits to the standard BOINC clients. I'm not on any of the teams mentioned and many of the others that were involved. As for the new credits being at pretty much the same level as the old ones, that was obviously going to happen - it's in-line with the standard. Why would they increase the credits to the level of 5.5? The users of 5.5 were/are a minority. You've taken this as the project favouring those with opioions different to yourselves, but I can't see why you find this a problem!

I can't understand the suggestion that everyone could have used 5.5 as a solution. I know you've stated that your were (are?) all for a fair credit system, but there are others who consistently state that this should have / could have been used! 5.5 isn't aligned to Rosetta throughput on a CPU. For example, from the new credit system we've seen that a larger cache has a very positive impact on productivity - my Duron has suffered! Maybe it's true that it levelled the playing field between the P4s and the AMDs but not in any way accurately, and what about all the other configs like Linux etc.

You will get the many thousands to assist your project who know nothing of what is really going on and they will be content that they are helping science go forward when the reality is they are helping a project led by a man who doesn't care to take the time to understand the people who donate so much to him. My eyes have been opened to the type of person he really is.
His own statements say it, I don't have to

I've seen David Baker's comment about having gained thousands of needy children mentioned a few times and it has been taken out of context, but is spot on in any context! Read the threads - people are complaining about being called names and points systems etc! I'm fairly sure he

I've been involved in the debate(?) and so am not innocent, but I would never consider my contribution worthy of being able to demand time from anyone on the project. If your eyes have been opened to the type of person he might be, then the only valid conculsions you could come to are 'busy' and 'shocked'...

The entire project team have the potential to have a massive impact on the future of biochemistry, including medicine. That is infintely more important than catering to any of our requests and desires - the ONLY consideration should be what will get them the resources they need to do the work. We can help that by not taking, requesting or expecting their time beyond what is absolutely necessary.

cheers
Danny

---

Actually my post, in my opinion, was about the apparent lack of respect the project has for me and my Macs because in my opinion, the client was not written with interest in regard to the PPC and it's features. I do not agree that it is acceptable to provide a client that is so flawed that it will make me waste 2 thirds of my processor time and investment thinking I am doing something.

If the client can not produce more that 1 third usage of a Mac PPC processor Don't distribute it.
I have seen other projects that do not have a Mac client.

Those projects don't put out a client that barely works and say here you run this.

Dave

It isn't about respect, it's about the suitability and capability of the processor for the specific code and, possibly more importantly, the ability of the compiler. You suggest they shouldn't release a client for the PPC, but at least this way they've given you the option.

It might be possible for extensions such as Altivec to be used, but then it might also be possible to run Rosetta on a GPU, or get another 20% from an Intel/Windows platform etc. There is a limited resource to do this work unfortunately.

P.S. where have you got the 33% figure from?

---

It might be possible for extensions such as Altivec to be used, but then it might also be possible to run Rosetta on a GPU, or get another 20% from an Intel/Windows platform etc. There is a limited resource to do this work unfortunately.

P.S. where have you got the 30% figure from?

I got the 30% from actual results turned in on my DP 2.5 G5 that was at 1,100 for 16 hours and went down to 220 for the same 16 hours with the new credit system, that is both processors not each.

So if I actually use a calculator I stand corrected make that 20%.

Feel better?

---

I got the 30% from actual results turned in on my DP 2.5 G5 that was at 1,100 for 16 hours and went down to 220 for the same 16 hours with the new credit system, that is both processors not each.

So if I actually use a calculator I stand corrected make that 20%.

Feel better?

Might it be a case that your original figures for the first 16 hours might have been excessive, at least to some extent? 1,100 is a very high score - I'd expect it to be around half that for a dual CPU machine at that speed. 220 seems low though - I'd expect a RAC of somewhere around 500-550 for a 24/7 AMD/Intel (non-P4) machine under low load.

I haven't seen a comparison of PPC throughput on Rosetta compared to AMD and Intel CPUs but that's what you need to compare isn't it?

---

Dave:
Please consider posting in this thread and post your old/new credits in this thread if that's the thread that was comparing the P4s, Athlons, and perhaps the Macs.

What's different between your Mac and David Kim's? I've seen mention of his system under the new credit system getting around 0.54? of the old credit system.

---

Rosetta Moderator: Mod.Tymbrimi
ROSETTA@home FAQ
Moderator Contact

Dave:
Please consider posting in this thread and post your old/new credits in this thread if that's the thread that was comparing the P4s, Athlons, and perhaps the Macs.

What's different between your Mac and David Kim's? I've seen mention of his system under the new credit system getting around 0.54? of the old credit system.

I usually consider a question and answer thread in a forum as the correct place to post it.
It was agreed that the first post I made was correctly in this forum.
I have responded to asked questions or comments about what I said in this forum so this is still appropriate.
If I posted the answer to this or another post in this forum on a different forum, Who would know?
How would you find it?
And I will go out on a limb here and answer your question even though it is off topic, sort of, I don't know anything about David Kim's computer so I can not comment further.

---