|
Bill Allombert on Wed, 20 Mar 2024 23:59:43 +0100
|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|
Re: How to efficiently count Proth primes with GP parfor?
|
- To: pari-users@pari.math.u-bordeaux.fr
- Subject: Re: How to efficiently count Proth primes with GP parfor?
- From: Bill Allombert <Bill.Allombert@math.u-bordeaux.fr>
- Date: Wed, 20 Mar 2024 23:59:34 +0100
- Arc-authentication-results: i=1; smail; arc=none
- Arc-message-signature: i=1; a=rsa-sha256; d=math.u-bordeaux.fr; s=openarc; t=1710975579; c=relaxed/relaxed; bh=Ze2i+Hy+OBu6y42QYZ/gpMnGtPIriM3dg9ohYqa6Esc=; h=DKIM-Signature:Date:From:To:Subject:Message-ID:Mail-Followup-To: References:MIME-Version:Content-Type:Content-Disposition: In-Reply-To; b=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
- Arc-seal: i=1; a=rsa-sha256; d=math.u-bordeaux.fr; s=openarc; t=1710975579; cv=none; b=Jrk9Wdum6uLFs27JHSp5cQKjkSUYA/kRIgJrCMcUGVHrjoUqy9k0WaWHmX2MO+K/MdUjv95g8Nc4A2X1P9yb2zLL443cbxKuA4U4ffPER9fAUDDbDAiktHUgf0bPhKhGs5HvRdPLprLDY9MjZK2DSzRdJmtQNuwqVFZvPLCWohLdcw8mHDKS4XEFOhPrHm7aTE/716T/E2xF6BKrk0w8rzxwvdP9FiZkLBLQORUPvNXpewAuXz7gRQazn4YeurvHdh2/EpZEu1/6gRQ2PaiMycBhq6/ii3B+TucUAep62qCHf2FUDHNpfRg2oKAkupIbV7NgaJEC4UN4O7G13/q1scrg1ROLtR48DNV/mnyiB9g8VdOdSD60+RI0PkTBNclHUmmyjYBP2Qd1jTU8bXaqWUfKVqWzQ7700SuenGr++io4k3FFXiWEyfb1aRwo4X9v7pWexvawVDUMS/lYgqe/LpzL49rJ5kv3sG+07vBzgps3563/0BbaAfUNDk+uGkIE167OB6zHzZEzERnst5WCv+0vw1BdpLVCvGsi9KxqNPhTV7mQXU1oiNrOYzT5Jykfx9GUW0mxGbYgz7mjNsShra0yTAO8qSqdGMpSr9g1RNd37mdknMVpEO3URSfiaJxEgAUxeSbl/TBUADZpH71WSGQN2XRkXZ0ObIhPuvoQtg8=
- Authentication-results: smail; dmarc=none header.from=math.u-bordeaux.fr
- Authentication-results: smail; arc=none
- Delivery-date: Wed, 20 Mar 2024 23:59:43 +0100
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=math.u-bordeaux.fr; s=2022; t=1710975579; bh=Ze2i+Hy+OBu6y42QYZ/gpMnGtPIriM3dg9ohYqa6Esc=; h=Date:From:To:Subject:References:In-Reply-To:From; b=uwuF0Q7mFEjWusox3KMDX055qtAZHkK0SCThHfpgDTre161PxUzNGQYJbUSqHF4o8 Oe6Fw/MoG7N0DvDnjMEwnE+sB/7HF64cxe+rAWnTwQDgNPMYsij7S2H4KP531RX1CW vvO9+gcPBQ3Ssb2RC32DknHVBSv3h7itC7UthdC22GxnIxBXp5CvzOGRnxRkTIbh1A M3YRXxldvDXb/78N4YIVda6X+6pm52/lzY0U/d4DoMjeDZzjK19Oizb4TPdpXTrgl+ PCqIGPAhtU3yXakGo6zUjGZP7rVHDB8IjtouGCTjczx9e9vmdCIgQmJVemFSX40JtW 7xCVSKJsrbEVwfLgSLnZ0u1/B9AWVmH2kQN+wSvHRJa0TOVCFSY4teckYOox0tjzo5 CniQyZFzcAasWjsOYL/wISk8v9UIoqXpD4Z5eenIpICN6e16EiOhsBGHe18HK6Zttw fL8ziQ0XOxmJ5ekSwOp9tkSVyI6PRfM8ZSdH/rOqZzy2FSZ3LttCn9bG8DPEjKS6NY HZqjBq3TzUKiMZLCLkOUuLt96GIDqYz8hLRT7la+owl+n9HiySfSpfmzB/1qgDrgwF Aab52DXmSlIZtVdZo0jMYRw4459gx9iLkpB7NVzCSegGE8l5eYgtfiPDxXY4/Rw9MB XS1hHFUk+aizY2nbZYP0tIAU=
- In-reply-to: <3944d2b195d4b2403b473ce5ecbbe4f7@stamm-wilbrandt.de>
- Mail-followup-to: pari-users@pari.math.u-bordeaux.fr
- References: <78df8c6fb5e60216469075d862f79735@stamm-wilbrandt.de> <55fecf62eb12a6076cc03e3ff0c06bce@stamm-wilbrandt.de> <Zfq57ke3bDo3Yx7M@seventeen> <f716073db48b3b39381af4041762e422@stamm-wilbrandt.de> <ZfrtMSOWpVYKn2F0@seventeen> <0b08544f1ae0f0459d0a589258f9489f@stamm-wilbrandt.de> <3944d2b195d4b2403b473ce5ecbbe4f7@stamm-wilbrandt.de>
On Wed, Mar 20, 2024 at 11:20:50PM +0100, hermann@stamm-wilbrandt.de wrote:
> >
> While using Karim's isProth2() I improved Bill's parallel work distribution
> scheme.
>
> Instead of blocks of size B=n\nbt now block are of size B=n\nbt\if(l,l,1).
> Since distribution of primes as well as distribution of Proth primes is not
> the same for lower half ad higher half, making each of the 32 hardware
> threaas work l smaller but more evenly distributed work areas.
Something else you can try is to increase nbt, to have smaller blocks.
Cheers,
Bill.