Great news!


b.t.w.: the Tflop counter I use in my signature (bottom line) is now at 388 Tflops. It's the average of the last 30 days minus the 2 highest days en the 2 lowest days. Shouldn't take long for it to reach the 400 mark too.

Excellent !

In time it will be made so... from January to May to add 17 TFL was tough... the power per CPU year is the big driver presently... nearing 1.3 TFL per CPU year contributed.

ok, so how do I tell how many of what type of flops I have crunching...
my CPU benchmark said for 8 cores
3040 floating point MIPS (whetstone) per CPU
9949 Integer MIPS (Dhrystone) per CPU
Ok, so using Dhrystone (that's what wiki was listing I think)
I'm doing
79592 (for a I7 920 OC to 3.3 ghz, they said 76,383 for a i7 965EE @ 3.2 ghz closest match I could find)
Interesting note, an I7 980x was rated at 147,600 @ 3.33
almost double with only 50% more cores... (I'm guessing memory cache and higher RAM might help?)

So 79592 Ima call it 80000 for simplicity sake.

I just tried to read up on this in wiki for last hour while typing this...

And I realize I dont know anything

I still can't figure out the difference between MIPS (million instructions per second)
and FLOPS (floating operations per second)
never mind how many FLOPS my machine does because BOINC reports it in MIPS...
Ok, MIPS was earlier simpler method of measurement compared to FLOPS.
Does MIPS as reported by BOINC correspond to FLOPS?
I'm doubtful as there are wet and dry versions.

By the way I do get that a terraflop is a trillion Flops so WCG
is doing 401 terraflops on average, that was what got me started
on this mind bender at 4am when I couldnt sleep. Now I can't sleep or think.
I apologize to all of you who read this drivel, just imagine what living
with my mind is like.
Here is my math someone please tell me if i'm in the ballpark
Assuming MIPS = FLOPS for this calculation
400,000,000,000,000 flops (or 400 terraflops) divided by 80,000,000,000= 5000 I7 920 @ 3.3 ghz (my machine)
Somehow that number seems extremely low, or with 500,000 (if i remember correctly) devices registered either they are REALLY slow old ones or they are devices that have for whatever reason been reregistered(reformating for example), blew up, or belonged to people that quit. Or more likely, my assumptions are wrong or my math is. Sorry if this is really stretching the topic, or if this has been addressed before. (I am going to take my points earned per day and points on WCG and try the math that way) Thanks for
any thought/help I can get.

OK so I told you I just came down (almost 3 hours ago now I'm taking my time writing and reading) cuz I couldnt sleep, well I just
reread my post for the umpteenth time and realized it says,
"3040 floating point MIPS (whetstone) per CPU
9949 Integer MIPS (Dhrystone) per CPU"
yes I really typed Floating point MIPS and never noticed it said that.
so should I be using whetstone for Flops (Floating point Operations Per Second)?
since they both say Floating in their names...
If so that would very roughly be like 16,000 I7 920s

EDIT: So I took points yesterday 286,194,142 divided by 22,000 (a little off the top scores of 24,000 for my I7)
and came up with the equivalent of 13,000 of my computers running. Sounds like I may have gotten lucky

Jack007,

FLOPS is not exactly what they are since the benchmarks are summed for FPOPS and IOPS and divided by around 480 to get an hourly credit potential. 100,000 of these credits also known as cobblestones are considered equal to 1 TFL sustained.

Note that Not allot, is per second, so the daily calculations need multiplication by 86,400 seconds in a day to find the total operations executed.

My Q6600 does about 1,800 credit per day. WCG totals about 40 million. Do that times 10 gives the Daily TFL of 22,000 quads doing 24/7. Of course many devices don't do that [the principle of distributed volunteer computing], so we recon about 200,000 devices are contributing something on a daily bases to WCG. Imagine if all would run 24/7 :D