The search up to 10^20

Message boards : News : The search up to 10^20

To post messages, you must log in.

AuthorMessage
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 521 - Posted: 30 Jun 2017, 16:11:33 UTC

It will be started tomorrow (July 1st)!

I've decided to speed up the launch of the next search due to performance issues with GPU version that everyone experiences now.

What will happen on July 1st:

1) "Amicable Numbers up to 10^20" application will start receiving tasks.
2) All GPU versions for the old application "Amicable Numbers up to 2^64" will be deprecated.
3) Credit points for both applications will be re-balanced to make sure that RAC doesn't change.

So all GPUs should automatically switch to the new application, and the old application will remain as CPU-only until it finishes.
ID: 521 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Henrik Nilsson

Send message
Joined: 27 Apr 17
Posts: 9
Credit: 10,986,936
RAC: 23,438
    
Message 522 - Posted: 30 Jun 2017, 17:22:13 UTC - in response to Message 521.  

>So all GPUs should automatically switch to the new application,
>and the old application will remain as CPU-only until it finishes.

But there will also be CPU tasks for the 10^20 application right from the start, correct?
ID: 522 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 523 - Posted: 30 Jun 2017, 17:34:05 UTC - in response to Message 522.  
Last modified: 30 Jun 2017, 17:36:14 UTC

But there will also be CPU tasks for the 10^20 application right from the start, correct?

Yes. You can see all application versions here: https://sech.me/boinc/Amicable/apps.php
ID: 523 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mmonnin

Send message
Joined: 19 Mar 17
Posts: 5
Credit: 33,586,785
RAC: 50
   
Message 529 - Posted: 30 Jun 2017, 23:53:15 UTC

Will there be more pairs to be found to count towards the badge?
ID: 529 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 530 - Posted: 1 Jul 2017, 6:41:07 UTC - in response to Message 529.  

Will there be more pairs to be found to count towards the badge?

There are ~1,750,000 undiscovered pairs in the range 2^64 ... 10^20. Since each pair is found by two different users usually, it means ~3,500,000 "to be found to count". But this will all happen over a long period of time (more than a year), because this search range is much larger than before.
ID: 530 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Su_Root@jisaku

Send message
Joined: 16 May 17
Posts: 3
Credit: 10,078,653
RAC: 0
  
Message 536 - Posted: 1 Jul 2017, 16:42:50 UTC

Thank you!
My GPU load is 100% and CPU load is only 2%
ID: 536 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
pancho

Send message
Joined: 4 Jun 17
Posts: 1
Credit: 17,310,658
RAC: 0
  
Message 537 - Posted: 1 Jul 2017, 21:19:47 UTC

Is there any way I could convince you to reduce the required VRAM back to 1.5gb? I've had to switch projects on a bunch of my boxes because I saw that they were no longer acceptable by only 375 mb or so....
ID: 537 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 544 - Posted: 2 Jul 2017, 8:44:14 UTC - in response to Message 537.  

Is there any way I could convince you to reduce the required VRAM back to 1.5gb? I've had to switch projects on a bunch of my boxes because I saw that they were no longer acceptable by only 375 mb or so....

I can reduce it, but the problem is that the GPU version will still try to use more than 1.5GB when it runs.
ID: 544 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
nomarski

Send message
Joined: 18 Apr 17
Posts: 1
Credit: 12,722,642
RAC: 0
  
Message 552 - Posted: 3 Jul 2017, 14:22:36 UTC

One more sad Boincer here. I have to give up my Amicable Numbers contribution because of the 2GB memory requirement. Is this because the computing becomes more complex now? Or can we expect this limit getting lower again somewhere in the future?
ID: 552 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 553 - Posted: 3 Jul 2017, 15:48:58 UTC - in response to Message 552.  

One more sad Boincer here. I have to give up my Amicable Numbers contribution because of the 2GB memory requirement. Is this because the computing becomes more complex now? Or can we expect this limit getting lower again somewhere in the future?

The algorithm has to store and use all prime numbers up to square root of N*2 where N is the search limit. After we switched to 10^20, this prime numbers table requires ~1208 MB alone, other internal buffers also increased in size a bit, so the program doesn't even fit in 1.5 GB anymore.
ID: 553 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Felix_M_

Send message
Joined: 17 Jun 17
Posts: 6
Credit: 5,387,608
RAC: 10,895
  
Message 613 - Posted: 16 Aug 2017, 10:04:41 UTC

i know its early for this question , but what will be after 10^64?
ID: 613 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 614 - Posted: 16 Aug 2017, 11:41:54 UTC - in response to Message 613.  

I haven't decided yet. The search up to 10^20 will take more than 4 years at current speed, so 10^21 will take more than 40 years which is just too much. So it all depends on how many volunteers will be running the project in 4 years from now.
ID: 614 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Felix_M_

Send message
Joined: 17 Jun 17
Posts: 6
Credit: 5,387,608
RAC: 10,895
  
Message 615 - Posted: 16 Aug 2017, 11:57:18 UTC

okay thank you
ID: 615 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Richard COLLINS

Send message
Joined: 23 Dec 17
Posts: 3
Credit: 1,025,429
RAC: 0
  
Message 703 - Posted: 4 Jan 2018, 12:03:30 UTC - in response to Message 614.  

When (GRC) goes from $0.06 USD to $0.18 USD in a month I can tell you that uses will be here very soon. When I say users I am talking about a lot more users than what you currently see here at grcpool, much more than what is here now. Currently 1 magnitude equals about 0.25 GRC/day @ $0.18 USD per GRC

I just swam in from a pool with a total hashrate of 43.93 MH/s on (XMR) Monero. To give you a feeling of how much that is lets look at an RX 470 GPU's hashrate of 600H/s when mining a cryptonight algorithm like XMR and hold that next to the pools total combined hashrate of 43.93MH/s. [43,930,000/600=73216] We can see that the pool has the GPU power of about 73,216 RX 470's. Then lets calculate how much GPU power is within the entire XMR network which is currently at 593.76MH/s for a sum of every XMR pool. [593,760,000/600=989600] That means the XMR network alone has a GPU equivalency of 989,600 RX 470 GPU's. That is only one of hundreds of coins people are mining and it is not the most popular either.

Most people mine what is immediately profitable and are they are not faithful to anything but profitability. Currently a single GPU can make $3 a day mining ETH or $2 a day mining XMR or $1 a day or less researching on BOINC. Once (GRC) is worth $0.50+ (March 2018 - Quote me on this one) you will see so many users here that the only problem to worry about then will be acquiring new WU's, keeping positive pool balances and mitigating SYN/UDP Floods from DDoS haters.

I like the MAG and RAC calculation system and how it keeps you faithful and committed. However this system currently will make a normal crypto miner quitting 24 hours after starting his preliminary test run when he see's only 1/30th of his actual potential since RAC is 30 day based (please correct me if I am wrong)

To test the waters I ran a 6GPU RX 470 rig nonstop on grcpool for 4 solid days which landed me in first place for daily credit (2,000,000+/day Amicable Numbers) and also I was user of the day by the third day. I was in first place generating 3x the daily credit as the 2nd place runner up was. Imagine what a whale with 4000 GPU's would do to this pool. Oh and yes they exist - search youtube for this "#EvolveWithUs - The Series / Official Trailer" or "My visit Inside Genesis Mining"

I am also going to start a new discussion with this same reply. Keep up the good work Sergei Chernykh.
ID: 703 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile [AF>Amis des Lapins] Naz

Send message
Joined: 24 Jan 17
Posts: 5
Credit: 1,180,639
RAC: 0
   
Message 704 - Posted: 16 Jan 2018, 14:45:08 UTC - in response to Message 614.  

I haven't decided yet. The search up to 10^20 will take more than 4 years at current speed, so 10^21 will take more than 40 years which is just too much. So it all depends on how many volunteers will be running the project in 4 years from now.


Why not run the 10 ^ 21 calculations in parallel?
ID: 704 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Kellen

Send message
Joined: 14 Nov 17
Posts: 10
Credit: 425,546,216
RAC: 2,441,949
    
Message 706 - Posted: 19 Jan 2018, 0:07:41 UTC - in response to Message 553.  

One more sad Boincer here. I have to give up my Amicable Numbers contribution because of the 2GB memory requirement. Is this because the computing becomes more complex now? Or can we expect this limit getting lower again somewhere in the future?

The algorithm has to store and use all prime numbers up to square root of N*2 where N is the search limit. After we switched to 10^20, this prime numbers table requires ~1208 MB alone, other internal buffers also increased in size a bit, so the program doesn't even fit in 1.5 GB anymore.


What would the RAM requirement be for the search to 10^21? I would guess that it wouldn't quite be a 10x jump to 15GB, but it could still be quite substantial.
ID: 706 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sergei Chernykh
Project administrator
Project developer

Send message
Joined: 5 Jan 17
Posts: 340
Credit: 72,451,573
RAC: 0
   
Message 707 - Posted: 19 Jan 2018, 6:26:30 UTC - in response to Message 706.  

What would the RAM requirement be for the search to 10^21? I would guess that it wouldn't quite be a 10x jump to 15GB, but it could still be quite substantial.

The memory required is O(sqrt(N)), so it will not be a 10x jump:
CPU version would require ~1400 MB of memory.
GPU version would require ~3700 MB of GPU memory.
ID: 707 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Kellen

Send message
Joined: 14 Nov 17
Posts: 10
Credit: 425,546,216
RAC: 2,441,949
    
Message 709 - Posted: 20 Jan 2018, 2:35:53 UTC - in response to Message 707.  

What would the RAM requirement be for the search to 10^21? I would guess that it wouldn't quite be a 10x jump to 15GB, but it could still be quite substantial.

The memory required is O(sqrt(N)), so it will not be a 10x jump:
CPU version would require ~1400 MB of memory.
GPU version would require ~3700 MB of GPU memory.

Great! Thank you for the quick reply! That is better than I thought. 3700MB for GPU would still leave a large portion of GPUs capable of participating and only 1400MB for CPU is quite low so I doubt anyone would be left out of the search up to 10^21 if it occurs.
ID: 709 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : News : The search up to 10^20


©2018 Sergei Chernykh