Multi Core Pi v2.000

[Sable]

Seems a bit more consistent this time.

2500k @ 4.7ghz

 

[John Tauwhare]

4.8 22.9
4.9 22.4
5.0 22.2
5.1 21.5
5.2 21.1
5.3 20.7
5.4 20.5
Sort of linear.

 

[Prey2big]

Ran it on one of my old machines just for fun. Mobo is an Abit IC7 with the i875P chipset.

 

[zebrax2]

 

[crashtech]

My old junk:

 

[Centauri]

http://oi48.tinypic.com/2j5dpa0.jpg
Ran it on one of my old machines just for fun. Mobo is an Abit IC7 with the i875P chipset.

If ever you needed an illustration of how much IPC has improved over the past decade... Crazy.

 

[YBS1]

 

[crashtech]

It's not terribly significant, but my E5472s are running at stock speed. I do not know why 3003 was reported, a whopping 1/10% overclock...

Oh, yeah, and there are two of them. I did not see a column for number of CPUs.

 

[bgt]

 

[Idontcare]

Wait, didn't we do this once and it was concluded the benchmark was pants?

It appears to be pants for AMD chips, Intel chips seem to behave as expected.

Note the author has recompiled and released the app under the same link without any external indication of having done so.

It now runs as rev 2.101, whereas everyone else generated rev 2.000 results.

You can see the difference in the icon located in the lower-left corner.

edit: wow, he even ninja-edited his own thread to replace the download link with the new file while keeping the download counter untouched and makes absolutely no reference to having deprecated his original software with the newer version. How on earth this guy and his benchmark are allowed to be HWBOT accredited is beyond me. Complete BS.

 

[Prey2big]

Ran it again, this time on my main rig with 1 core (+HT) at similar clockspeed as my P4.

 

[mojothehut]

Does the Phenom II just suck at this test? Some Intel chips, a 1000mhz slower finished like 20 seconds faster :hmm:
Anyway this a 975 @ 4.0ghz

 

[Idontcare]

It appears to be pants for AMD chips, Intel chips seem to behave as expected.



Note the author has recompiled and released the app under the same link without any external indication of having done so.

It now runs as rev 2.101, whereas everyone else generated rev 2.000 results.

You can see the difference in the icon located in the lower-left corner.

edit: wow, he even ninja-edited his own thread to replace the download link with the new file while keeping the download counter untouched and makes absolutely no reference to having deprecated his original software with the newer version. How on earth this guy and his benchmark are allowed to be HWBOT accredited is beyond me. Complete BS.

Ran it again, this time on my main rig with 1 core (+HT) at similar clockspeed as my P4.

LOL, he's now changed it to rev 2.200

 

[Edgemeal]

LOL, he's now changed it to rev 2.200

Its possible the benchmarking code itself hasn't changed and only other parts like uploading the result and the visual appearance were updated, but without notes/history its any ones guess.

FWIW, I just ran 2.1 and 2.2 a few times and getting the same time as I was in 2.0.

 

[grimpr]

Guys why are you still wasting your time in this crap? Sending an email request to y-crunchers author to release the new version asap would be more beneficial.

 

[Vectronic]

2.101 He made the output/results encrypted
2.200 He removed the digit slider so it's always 80,000

As seen in this small thread.

Apparently it went from MultiCore PI, to MultiCore Prime, back to MultiCore PI, all of them poorly coded bloatbombs.

 

[ThatsABigOne]

 

[Vectronic]

I honestly don't know what this application does... simplification, but I tried this:

For n As Integer = 0 To 8000000
d += Math.Pow(16, -n) * ((4 / (8 * n + 1)) - (2 / (8 * n + 4)) - (1 / (8 * n + 5)) - (1 / (8 * n + 6)))
Next

Takes about 3 seconds to calculate that... obviously d is limited to: 3.14159265358978~ but I don't see how it's taking 35 seconds for it to calculate 80,000 in "MultiCore PI" and 3 seconds for 8,000,000 iterations in mine.

Something else I discovered, this equation is s**t... if you wanted to do it by hand, or you had some specially built 2048bit CPU awesome... otherwise it's not very accurate.

 

[Puppies04]

Laptop in my sig.

 

[Idontcare]

I honestly don't know what this application does... simplification, but I tried this:

Takes about 3 seconds to calculate that... obviously d is limited to: 3.14159265358978~ but I don't see how it's taking 35 seconds for it to calculate 80,000 in "MultiCore PI" and 3 seconds for 8,000,000 iterations in mine.

Something else I discovered, this equation is s**t... if you wanted to do it by hand, or you had some specially built 2048bit CPU awesome... otherwise it's not very accurate.

True. On my 2.3GHz i5-2410M laptop (2C/4T) running a single-threaded calculation in mathematica using the exact same formula that the Multi-core Pi uses, it takes 1.202 seconds to compute 80k digits.

 

[sstephan]

mine

 

[Centauri]

A10, represent!

 

[Vectronic]

mine

Do you happen to have a copy of MS Office installed?... if so, jump over to here.