mersenneforum.org  

Go Back   mersenneforum.org > Search Forums
Reload this Page Search Results
Register FAQ Search Today's Posts Mark Forums Read

Showing results 1 to 20 of 20
Search took 0.01 seconds. Search: Posts Made By: kwalisch
Forum: Computer Science & Computational Number Theory 2020-10-29, 07:54
Replies: 34
Views: 18,339
Posted By kwalisch
We have successfully verified PrimePi(1e28)

We have successfully verified PrimePi(1e28). I mistakenly posted the announcement in other thread: https://mersenneforum.org/showthread.php?p=561248#post561248

Since we now have computed already...
Forum: Computer Science & Computational Number Theory 2020-10-28, 07:54
Replies: 50
Views: 17,412
Posted By kwalisch
That's actually a very good guess! I paid: ...

That's actually a very good guess!

I paid:

AMD EPYC 7642: 5500 EUR
512 GiB DDR4 ECC 3200 MHz: 3100 EUR
Supermicro A+ Server AS-4023S-TRT: 1800 EUR
Fractal Design Define 7 XL: 200 EUR
1000...
Forum: Computer Science & Computational Number Theory 2020-10-27, 16:38
Replies: 50
Views: 17,412
Posted By kwalisch
We have successfully verified PrimePi(1e28)

I am happy to announce that David and myself have successfully verified PrimePi(1e28) = 157589269275973410412739598. This time the computation took only 12.8 CPU core years (only physical CPU cores...
Forum: Computer Science & Computational Number Theory 2020-08-30, 20:38
Replies: 34
Views: 18,339
Posted By kwalisch
Implementing primecount was/is a lot of work,...

Implementing primecount was/is a lot of work, when I started the project back in 2013 you were a tremendous help to myself. At that time both your and my project included the only advanced prime...
Forum: Computer Science & Computational Number Theory 2020-05-06, 07:21
Replies: 34
Views: 18,339
Posted By kwalisch
primecount is still being actively used to...

primecount is still being actively used to compute new records by David Baugh and myself. primecount has been massively improved since our last pi(10^27) announcement i.e. I have now implemented...
Forum: Software 2019-09-22, 09:59
Replies: 543
Views: 59,113
Posted By kwalisch
This is not true, vector<bool> is special, it may...

This is not true, vector<bool> is special, it may be implemented as a bit vector, see the documentation here: https://en.cppreference.com/w/cpp/container/vector_bool. At least 5 years ago it was...
Forum: Computer Science & Computational Number Theory 2019-09-01, 08:47
Replies: 6
Views: 2,845
Posted By kwalisch
Yes, something like that exists in primesieve and...

Yes, something like that exists in primesieve and it's called pre-sieving of small primes. At startup primesieve initializes a small buffer of size p1*p2*p3*pn and removes the multiples of the first...
Forum: Number Theory Discussion Group 2018-03-12, 22:25
Replies: 43
Views: 7,340
Posted By kwalisch
I don't get it. My understanding is that there...

I don't get it. My understanding is that there will be exactly one loop iteration per prime in the inner while loop (that generates primes).

Removing the multiples of 2, 3 and 5 from the intervals...
Forum: Number Theory Discussion Group 2018-03-10, 18:38
Replies: 43
Views: 7,340
Posted By kwalisch
The inner while loop that generates primes...

The inner while loop that generates primes compiles to 7 - 9 assembly instructions on x64. Each 64-bit word corresponds to an interval of size 240 and using the prime number theorem we can calculate...
Forum: Number Theory Discussion Group 2018-03-09, 23:17
Replies: 43
Views: 7,340
Posted By kwalisch
It's great to hear that it works! The good thing...

It's great to hear that it works! The good thing about having 64-bit words where the bits are arranged sequentially is that there exists a simple & fast algorithm to iterate over the 1 bits and...
Forum: Number Theory Discussion Group 2018-03-09, 13:53
Replies: 43
Views: 7,340
Posted By kwalisch
Faster sequential prime number generation in yafu

I have been thinking whether it was possible to speed up sequential prime number generation in yafu for a long time. And I always thought that maybe it would be a good idea to interleave the bits...
Forum: Number Theory Discussion Group 2018-01-28, 22:18
Replies: 43
Views: 7,340
Posted By kwalisch
Yes, it's true, yafu is faster than primesieve...

Yes, it's true, yafu is faster than primesieve for counting primes in the medium range e.g. [10^11, 10^16].

But this is by purpose, primesieve has been optimized for generating primes in...
Forum: Science & Technology 2018-01-04, 23:06
Replies: 42
Views: 17,337
Posted By kwalisch
According to the latest Google Security Blog post...

According to the latest Google Security Blog post (https://security.googleblog.com/2018/01/more-details-about-mitigations-for-cpu_4.html) the “Retpoline” software mitigation technique "may be applied...
Forum: Science & Technology 2018-01-04, 18:50
Replies: 42
Views: 17,337
Posted By kwalisch
Maybe it is not as bad as expected, according to...

Maybe it is not as bad as expected, according to the patch description LLVM/Clang will add a new compiler flag -mretpoline to generate secure code. I sincerely hope LLVM/Clang and GCC won't enable...
Forum: Science & Technology 2018-01-04, 18:36
Replies: 42
Views: 17,337
Posted By kwalisch
Angry Initially I also thought that most HPC software...

Initially I also thought that most HPC software was not affected because HPC programs contain mostly highly optimized number crunching loops with very few system interrupts. But then I came across...
Forum: Hobbies 2017-08-14, 21:07
Replies: 10
Views: 7,325
Posted By kwalisch
Yes, I have a Raspberry pi 3 and I have installed...

Yes, I have a Raspberry pi 3 and I have installed openSUSE 64-bit (AArch64) (https://en.opensuse.org/HCL:Raspberry_Pi3) on it in order to test the ARM Neon popcnt...
Forum: Computer Science & Computational Number Theory 2016-06-01, 21:07
Replies: 34
Views: 18,339
Posted By kwalisch
Thanks, I actually created an account on OEIS...

Thanks, I actually created an account on OEIS earlier today in order to submit it. But you have been faster. So I hope I can use my OEIS account in near future for submitting new prime sum records ;-)
Forum: Computer Science & Computational Number Theory 2016-06-01, 07:24
Replies: 34
Views: 18,339
Posted By kwalisch
pi(10^27) verified!

After roughly 5 months of computation David Baugh and myself have verified pi(10^27):

pi(10^27) = 16,352,460,426,841,680,446,427,399

This time the computation took 20.35 CPU core years, this is...
Forum: Computer Science & Computational Number Theory 2015-11-25, 12:46
Replies: 50
Views: 17,412
Posted By kwalisch
Thanks David! Are few more details about the...

Thanks David!

Are few more details about the computation:

pi(10^26) = 1,699,246,750,872,437,141,327,603

The computation took 34 days on David's dual socket server (36 CPU
cores, Intel Xeon...
Forum: Computer Science & Computational Number Theory 2015-09-06, 19:58
Replies: 34
Views: 18,339
Posted By kwalisch
New confirmed pi(10^27), pi(10^28) prime counting function records

I am happy to announce that David Baugh and I have computed the number
of primes below 10^27, the result is:

pi(10^27) = 16,352,460,426,841,680,446,427,399

The computation was performed using...
Showing results 1 to 20 of 20

 
All times are UTC. The time now is 00:09.

Sun Mar 7 00:09:41 UTC 2021 up 93 days, 20:20, 0 users, load averages: 1.84, 2.18, 2.08

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.