RMA.NET - Page 5

Flatlander · 2005-10-20, 11:55

Just niceties. (Working from memory here):

14) Automatically detect tick next to SSE2/Non SSE2.

15) Remember visible/hidden status of llr/newpgen between timers.

16) View current n without opening llr. (Have 'Current/n/k xxxx' below 'Has xxxx Digits')

17) Change 'File/Parent Directory' to 'File/View Running Directory' or on new 'View' Menu.

18) Ability to empty list of primes found. ('Are You Sure... yes/no' dialog. Hidden backup?)

19) Change buttons on 'Search for $100,000 prime?' dialog to yes/no.

20) Stick 'About' in 'Help' menu.

21) Change 'Flatlander' to 'Chris Chatfield' in credits.

(How else am I going to be famous?)

22) Make 'Add New Mersenne' button work. (Ability to delete incorrectly entered ones?)

23) Don't shrink expanded rma form when llr/newpgen closes.

24) - 82)...?

2005-10-20, 13:15

Quote:

14) Automatically detect tick next to SSE2/Non SSE2.

LLR usually auto detects SSE2 i think.
I made the option to shut it down, or force it.
If neither are checked in RMA, it should auto-detect by default.
I found out the hard way that LLR, may change it's default behaviour, so it's better to be neutrally checked by default.
This goes for most other applications, where the version may change.

I'll make llr test options changable during runtime, since there is only a small chance of llr error.

Quote:

24) - 82)...?

Not sure what 82?

The rest sound doable!
Thanks Chris.
Shane F.

2005-10-20, 13:55

I've been running a large RMA.NET test for about 36.85 hours now.
k= 2145
n = 250000 thru 1250000
One million candidates of fixed k, RMAing.

The timer has increased several times, and is now at about 8 1/2 hours.
NewPGen has been the dominant process the entire time.
Eliminating about 1650 candidates within this last time period.
Testing between 1-2 candidates within that same time period.

So the handicap is applied to llr, by multipying process %,
which is about 99, by the test completion count of 2. Not enough yet to LLR.
The timer increases so far, are due to LLR, but will increase again as NewPGen starts to slowly die out. Then increase again when the LLR tests get large with n.(since k is fixed) This may repeat, depending on size, speed etc.

At worst I would expect a very small error margin, relatively speaking.
1/2 timer. Doing this manually often forces one to over-sieve by much more. Some call that sensible sieving.
It looks like RMA may start testing fairly soon since the time will be equal between newPgen, and LLR's current lower range, of the large file.

Flatlander · 2005-10-20, 14:00

24) Is a joke. Suggestions 24 to 82. Just to keep you busy.

23) Correction. I think llr/newpgen disappear when resizing rma.

Flatlander · 2005-10-20, 14:05

Bug?

RMA reports 3 primes found but the first one is missing in the pop-up list.
3 primes are in prime.txt

2005-10-20, 14:06

Got me, although I suspected your were kidding.

We've overlapped on the Re-size events. Cool, I'll get that.

2005-10-20, 14:09

Ah It doesn't want me to pre-initialize the Readline string.
Another MS bug, and/or bad example.

Not hard to fix, I did'nt need speed here anyway.

Flatlander · 2005-10-20, 14:19

16) And/or show current k/n in bubble help at icon?

Is the 'Credit' box actually used at the database?

Flatlander · 2005-10-20, 14:40

re. 1) If you can't get ability to run other llr/newpgen they will need a chance to save their position before being forced to close. (Or dialog to user "Please close all llr/newpgen now running." Buttons - "I have/Cancel rma.net"

btw your k=2145 has primes <=214994

2005-10-20, 15:10

Quote:

e. 1) If you can't get ability to run other llr/newpgen they will need a chance to save their position before being forced to close. (Or dialog to user "Please close all llr/newpgen now running." Buttons - "I have/Cancel rma.net"

It should be easier than I thought to run other llr/newpgen.
I'll need a small mental break before 0.9 is completed.
The close box must not interfere modally with Shutdown.
The user may shut down, and RMA has to respond immediately.
Position placement is no task, so I can include that.

Quote:

btw your k=2145 has primes <=214994

Thanks, I am just going over the range as a large example for all.
[ edited ]

Quote:

Flatlander 16) And/or show current k/n in bubble help at icon?

That's what I was thinking, like llr does.

Quote:

Is the 'Credit' box actually used at the database?

Yes the credit number reflects, the actual number of string names given credit.

Flatlander · 2005-10-20, 15:14

24) Statistical data form. 'Change List of Riesel Primes' to 'List of Riesel/Proth Primes'. And change the list to a format ready for copy/paste into 'submit primes' box at database.

I think you deserve a break before 0.9! I will start 0.8 on my other PC tomorrow.

2005-10-20, 11:55	#45
Flatlander I quite division it "Chris" Feb 2005 England 31×67 Posts	Just niceties. (Working from memory here): 14) Automatically detect tick next to SSE2/Non SSE2. 15) Remember visible/hidden status of llr/newpgen between timers. 16) View current n without opening llr. (Have 'Current/n/k xxxx' below 'Has xxxx Digits') 17) Change 'File/Parent Directory' to 'File/View Running Directory' or on new 'View' Menu. 18) Ability to empty list of primes found. ('Are You Sure... yes/no' dialog. Hidden backup?) 19) Change buttons on 'Search for $100,000 prime?' dialog to yes/no. 20) Stick 'About' in 'Help' menu. 21) Change 'Flatlander' to 'Chris Chatfield' in credits. (How else am I going to be famous?) 22) Make 'Add New Mersenne' button work. (Ability to delete incorrectly entered ones?) 23) Don't shrink expanded rma form when llr/newpgen closes. 24) - 82)...? Last fiddled with by Flatlander on 2005-10-20 at 12:07

2005-10-20, 14:19	#52
Flatlander I quite division it "Chris" Feb 2005 England 31×67 Posts	16) And/or show current k/n in bubble help at icon? Is the 'Credit' box actually used at the database? Last fiddled with by Flatlander on 2005-10-20 at 14:32

2005-10-20, 14:40	#53
Flatlander I quite division it "Chris" Feb 2005 England 100000011101₂ Posts	re. 1) If you can't get ability to run other llr/newpgen they will need a chance to save their position before being forced to close. (Or dialog to user "Please close all llr/newpgen now running." Buttons - "I have/Cancel rma.net" btw your k=2145 has primes <=214994 Last fiddled with by Flatlander on 2005-10-20 at 14:48

2005-10-20, 15:14	#55
Flatlander I quite division it "Chris" Feb 2005 England 81D₁₆ Posts	24) Statistical data form. 'Change List of Riesel Primes' to 'List of Riesel/Proth Primes'. And change the list to a format ready for copy/paste into 'submit primes' box at database. I think you deserve a break before 0.9! I will start 0.8 on my other PC tomorrow. Last fiddled with by Flatlander on 2005-10-20 at 15:21

2005-10-20, 13:55	#47
TTn 2·3²·13·17 Posts	I've been running a large RMA.NET test for about 36.85 hours now. k= 2145 n = 250000 thru 1250000 One million candidates of fixed k, RMAing. The timer has increased several times, and is now at about 8 1/2 hours. NewPGen has been the dominant process the entire time. Eliminating about 1650 candidates within this last time period. Testing between 1-2 candidates within that same time period. So the handicap is applied to llr, by multipying process %, which is about 99, by the test completion count of 2. Not enough yet to LLR. The timer increases so far, are due to LLR, but will increase again as NewPGen starts to slowly die out. Then increase again when the LLR tests get large with n.(since k is fixed) This may repeat, depending on size, speed etc. At worst I would expect a very small error margin, relatively speaking. 1/2 timer. Doing this manually often forces one to over-sieve by much more. Some call that sensible sieving. It looks like RMA may start testing fairly soon since the time will be equal between newPgen, and LLR's current lower range, of the large file.

2005-10-20, 14:00	#48
Flatlander I quite division it "Chris" Feb 2005 England 31·67 Posts	24) Is a joke. Suggestions 24 to 82. Just to keep you busy. 23) Correction. I think llr/newpgen disappear when resizing rma.

2005-10-20, 14:05	#49
Flatlander I quite division it "Chris" Feb 2005 England 31·67 Posts	Bug? RMA reports 3 primes found but the first one is missing in the pop-up list. 3 primes are in prime.txt

2005-10-20, 14:06	#50
TTn 9376₁₀ Posts	Got me, although I suspected your were kidding. We've overlapped on the Re-size events. Cool, I'll get that.

2005-10-20, 14:09	#51
TTn 1891₁₆ Posts	Ah It doesn't want me to pre-initialize the Readline string. Another MS bug, and/or bad example. Not hard to fix, I did'nt need speed here anyway.