mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Riesel Prime Search (https://www.mersenneforum.org/forumdisplay.php?f=59)
-   -   Time per iteration (https://www.mersenneforum.org/showthread.php?t=8683)

em99010pepe 2007-07-13 20:18

Time per iteration
 
Please post your benches.

k=15 for n=1.6M

4.6 ms on an AMD 64 3000+ (2GHz)
3.2 ms on an INTEL P4 3.0GHz
3.2 ms on a Core 2 Duo T5500 (1.66 GHz)

gd_barnes 2007-07-13 20:44

[quote=em99010pepe;110305]Please post your benches.

k=15 for n=1.6M

4.6 ms on an AMD 64 3000+ (2GHz)
3.2 ms on an INTEL P4 3.0GHz
3.2 ms on a Core 2 Duo T5500 (1.66 GHz)[/quote]


I'm not sure what all of the info. in LLR means for my computer info. so I'll show it all. It's an almost brand new dual-core Dell. I think it's the equivalent of your 3rd listing above. Here's the info.:

Intel(R) Core(TM)2 CPU T5500 @ 1.66GHz
CPU speed 1662.44 MHz
CPU features: RDTSC, CMOV, PREFETCH, MMX, SSE, SSE2

After a quick sieve to P=1G, I ran 15*2^1600001-1 through LLR for the first 200K iterations:
Time per iteration: 2.894 ms.

I'm glad to see that it topped your listed speeds! :smile: :wink:

Hope this is what you're after...


Gary

em99010pepe 2007-07-13 21:36

I want to buy a new machine so I hoping to see some more Core 2 Duo and AMD X2 benches.

Carlos

VBCurtis 2007-07-13 21:57

Roughly, an X2 performs at just more than a P4 of the same speed (in Ghz, not in rating). Your own benchmark shows that, and it holds across many speeds. Memory speed might change things by a few percentage points, but you didn't supply or ask for that kind of detail.

For Core2Duo and small FFT sizes, double the actual Ghz speed and subtract 10% for equivalent P4 speed. A 2.4Ghz C2D-6600 would run about the same as a P4-4300Mhz, for instance.

I have one of each of these chips at speeds different from yours, and my own estimates fit both my home-machine data and your supplied data; again, details such as memory speed or type, motherboard chipset, etc can alter this by a few percentage points. For deciding among architectures, it's good enough.
-Curtis

Flatlander 2007-07-13 22:25

Frying tonight.
 
Core2Duo E4300 overclocked from 1.8GHz to 3GHz (not recommended if you want your system to last):

1.62ms :smile:

But if I try to run both cores at that speed the temperature approaches 80C!
So the other core seives, this keeps the temperature to about 72C.

If I have both cores running LLR I have to drop to about 2.75GHz.

If the ambient goes above about 24C I have to slow things a little.

Asus P5B mobo, can't remember which RAM but I paid a bit extra for low latency.

VBCurtis 2007-07-14 07:44

I disagree with the "if you want your system to last" part. I have two systems that are pretty seriously overclocked: Celeron-566 that ran 3 yrs at 920 and 4 more yrs at 850. This chip has answered all of my "will it kill the chip?" concerns from the early overclocking days. It has run Prime95, LLR, or sieves since the day I built the system, and is currently sieving. I have restarted it once in the last 15 months (to add memory).

I also have a P4-1800@2500 that has run Prime95/LLR for 4ish years. Again, no sign of any damage or increased errors as time has passed.

That said, 72C sounds pretty hot! Amazing it can hold a 66% overclock under those conditions. How much voltage? I have a 6300 on a cheapo board; I'm trying to justify getting a gigabyte 965 and going for 3+Ghz.
-Curtis

Flatlander 2007-07-14 12:58

CPU voltage is automagically set by the (included with mobo) AISuite programme. Currently at 1.352.

It was concern over the high temps. that made me give the warning. I could crank the CPU fan up to full speed but I can't stand the noise and it would only lower the temp. by about 2C.
I am using an Arctic Cooling Freezer 7 Pro; very quite up to about 2000rpm, currently at (a slightly irritating) 2500, capable of about 2800. Good value at about £15 UK.

Above 3.06GHz and the system freezes, even with manual voltage adjustment.

(Maybe this thread should be in 'Hardware'?)

arminius 2007-08-30 08:54

I´ve 2 machines:
Intel(R) Core(TM)2 6400, 2x 2.13 GHz, Dell DM061, 2048 MB RAM, Windows Vista
Intel(R) P IV 530J, 3,0 GHz, Asrock 775V88+, 1024 MB, openSuse 10.2

For LLRing k=49 I got these results (time):
49*2^990613-1 -> 1646.541 sec Intel(R) P IV 530J (without HT)
49*2^1006861-1 -> 1356.785 sec Intel(R) Core(TM)2 6400 (1 core)

Why needs the P IV ~17% longer with ~1 Ghz more an a lower n? Maybe the architecture of the cpu?
Thanks in advance.


All times are UTC. The time now is 22:56.

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