mersenneforum.org  

Go Back   mersenneforum.org > New To GIMPS? Start Here! > Information & Answers

Reply
 
Thread Tools
Old 2013-05-19, 06:59   #1
Unregistered
 

2·2,647 Posts
Default Summary statistics meaning

I've been here but a short time, and managed by one way or another to resolve many uncertainties.
I now see the summary page saying 171% CPU results last 24 hours, and 2% CPU results last 90 days.
What do these figures mean?
Percentage of what?
  Reply With Quote
Old 2013-05-19, 09:06   #2
NBtarheel_33
 
NBtarheel_33's Avatar
 
"Nathan"
Jul 2008
Maryland, USA

21338 Posts
Default

Quote:
Originally Posted by Unregistered View Post
I've been here but a short time, and managed by one way or another to resolve many uncertainties.
I now see the summary page saying 171% CPU results last 24 hours, and 2% CPU results last 90 days.
What do these figures mean?
Percentage of what?
These gauges didn't make much sense to me either when they were first introduced. But here's what they mean.

Given the computers (and their specifications) that you have registered under your GIMPS account, there is a sum total of *possible* throughput of all of these machines. These gauges are informing you as to what percentage of this computing power your *actual* throughput represents.

For instance, just to keep it simple, let's say you have 20 Intel Core2Duo processors running at 2 GHz. Theoretically, then, you would be capable each day of producing 20 CPUs * 2 cores/CPU * 2 GHz/core * 1 day = 80 GHz-days. Now let's say that in the last 24 hours, you performed some work for GIMPS, and earned 12 GHz-days. Then your "last 24 hours" gauge would read 15%. Why? Because 12 GHz-days, your actual throughput, represents 15% of 80 GHz-days, which is your maximum theoretical throughput.

Similarly, continuing this example, we see that in a 90-day period, your computers would be capable of producing a maximum of 80 GHz-days/day * 90 days = 7,200 GHz-days in that 90-day span. Therefore, if in the last 90 days, you've done work totaling, say, 2,500 GHz-days, your "last 90 days" gauge would read ~35%, since 2,500 (actual throughput) is ~35% of 7,200 (maximal throughput).

You will notice, however, that these gauges only move discretely, whenever you are *credited* with GHz-days at the *completion* of an assignment, rather than moving continuously as your computers complete their assignments. This will cause the 24-hour gauge to jump around considerably, especially if you are doing LL work. The 90-day gauge moves a little more smoothly, but if you have a sudden surge or decrease in your throughput, or if you add or delete computers (and thereby increase or decrease your maximum theoretical throughput), you can see this gauge spiking or falling rapidly.

Using the above example, let's say you begin a new GIMPS account, are capable of 80 GHz-days per day, and it takes you 60 days to complete an LL assignment worth 150 GHz-days, and that is the only work that you undertake during that time frame. That means that for 59 days, your 24-hour gauge and 90-day gauge will both show 0%. But on the 60th day, whenever you turn in that completed LL, your 24-hour gauge will read 188% (150 = 188% of 80) and your 90-day gauge will read 2% (150 = 2% of 7,200). (Note that these gauges can be pegged well over 100%; in fact, this almost always will happen when you submit an LL or other multi-day assignment.)

Finally, note that these gauges really only make sense if you retire computers from your CPU list (under your account page) that are no longer active in GIMPS. For example, if you have the 20 Core2Duos discussed above, but only 12 of them contribute to GIMPS, your gauge readings will be too low, because your total results are still being compared to a theoretical maximum throughput for *20* computers, rather than that for just *12*.

By the way, if you happen to be interested, these "gauges" also exist for GIMPS as a whole: see http://www.mersenne.org/primenet/. Right now, for instance, GIMPS is operating at 106% CPU power for the day, 59% for the last week, and 28% for the last month. So, since GIMPS has produced ~2.2 million GHz-days in the last month, we can estimate the current maximum (theoretical) possible throughput of GIMPS: ~7.86 million GHz-days per 30 days, or 262,000 GHz-days per day.

Hope this explanation helps a little bit. Post again if you have other questions! Welcome to GIMPS, have fun, and good luck!!!
NBtarheel_33 is offline   Reply With Quote
Old 2013-05-19, 10:31   #3
Unregistered
 

2·32·5·7·11 Posts
Default

[QUOTE=NBtarheel_33;340931]These gauges didn't make much sense to me either when they were first introduced. But here's what they mean.

Thankyou for the explanation;
Now I better understand the limited usefulness of same.
  Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
meaning of regular flashes of hot lead wildrabbitt Hardware 8 2015-06-22 10:29
meaning of earth and globe science_man_88 Lounge 0 2010-10-22 19:48
Meaning and format of Phi, GF roger Miscellaneous Math 9 2007-01-21 06:33
what's the meaning of the numbers in the FACTORS.CMP wreck Software 2 2006-08-30 04:46
Sieve discussion Meaning of first/second pass, combined Citrix Prime Sierpinski Project 14 2005-12-31 19:39

All times are UTC. The time now is 12:52.


Mon Nov 28 12:52:25 UTC 2022 up 102 days, 10:20, 0 users, load averages: 1.40, 1.37, 1.24

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, 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.

≠ ± ∓ ÷ × · − √ ‰ ⊗ ⊕ ⊖ ⊘ ⊙ ≤ ≥ ≦ ≧ ≨ ≩ ≺ ≻ ≼ ≽ ⊏ ⊐ ⊑ ⊒ ² ³ °
∠ ∟ ° ≅ ~ ‖ ⟂ ⫛
≡ ≜ ≈ ∝ ∞ ≪ ≫ ⌊⌋ ⌈⌉ ∘ ∏ ∐ ∑ ∧ ∨ ∩ ∪ ⨀ ⊕ ⊗ 𝖕 𝖖 𝖗 ⊲ ⊳
∅ ∖ ∁ ↦ ↣ ∩ ∪ ⊆ ⊂ ⊄ ⊊ ⊇ ⊃ ⊅ ⊋ ⊖ ∈ ∉ ∋ ∌ ℕ ℤ ℚ ℝ ℂ ℵ ℶ ℷ ℸ 𝓟
¬ ∨ ∧ ⊕ → ← ⇒ ⇐ ⇔ ∀ ∃ ∄ ∴ ∵ ⊤ ⊥ ⊢ ⊨ ⫤ ⊣ … ⋯ ⋮ ⋰ ⋱
∫ ∬ ∭ ∮ ∯ ∰ ∇ ∆ δ ∂ ℱ ℒ ℓ
𝛢𝛼 𝛣𝛽 𝛤𝛾 𝛥𝛿 𝛦𝜀𝜖 𝛧𝜁 𝛨𝜂 𝛩𝜃𝜗 𝛪𝜄 𝛫𝜅 𝛬𝜆 𝛭𝜇 𝛮𝜈 𝛯𝜉 𝛰𝜊 𝛱𝜋 𝛲𝜌 𝛴𝜎𝜍 𝛵𝜏 𝛶𝜐 𝛷𝜙𝜑 𝛸𝜒 𝛹𝜓 𝛺𝜔