What do I score in pifast… it's going to be pretty slow I'm afraid, I only managed to get to
3.141592653 in around a second and then I ground to a halt….I've clearly slowed over time because I used to be able to go to about 15 decimal places in the same time.
Sadly I'm a few generations old now and I just had my meltdown patch to protect me from a midlife crisis :P
Oh.. you mean my pc, doesn't really matter it will be the same as any other stock i7 4790k in all likelihood.
When you do it people, run the hexus_pifast.bat file, not the .exe
Series computing time : 25.25
Division time : 3.24
InvSqrt time : 2.02
Final huge multiplication time : 1.32
—————————————————————
Total computation time : 31.93 seconds(~ 0.01 hours)
phenom II x6 1090t @stock 3.2ghz, 16GB 1600MHZ ram.
16.61 timing on my older and trusty i7-4790 running at 4.0 Ghz stock.
20.6 seconds.
Processor Intel(R) Core(TM) i5-3570K CPU @ 3.40GHz 7.6
Memory (RAM) 16.0 GB 7.8
Graphics AMD Radeon HD 6800 Series 7.8
Gaming graphics 4864 MB Total available graphics memory 7.8
Primary hard disk 158GB Free (233GB Total) 7.9
Rizen 2600 @ 3.8 gig
16 gig 3200 mhz ddr4
wow I am impressed about my 5 Y old CPU Intel® Core i7-4770 3.4GHz
Total computing time - 17.99
On my work computer: 23.45s
CPU: i5-7500T @ 2.70GHz
RAM: 16GB
Drive: 250 GB SSD
OS: Win10 Pro
Actually the bat file didn't work for me, I just inputted all the params manually.
Pleased with that. 15.6 secs
5960X @4.2, 32Gb@3000, Samsung SG561 on Win 10 pro
1800x stock
1800x 4ghz oc
Series computing time : 10.66
Division time : 1.13
InvSqrt time : 0.70
Final huge multiplication time : 0.47
—————————————————————
Total computation time : 13.00 seconds(~ 0.00 hours)
This is on a 7700k @ 5GHz
Ryzen 5 1600 @ 3.85ghz
Series computing time : 17.84
Division time : 1.96
InvSqrt time : 1.19
Final huge multiplication time : 0.83
—————————————————————
Total computation time : 21.88 seconds(~ 0.01 hours)
Lol
27.72 computing time
2.89 division time
1.80 Invsqrt time
1.20 final huge multiplication time
So a total of 33.71
Xeon X5645 @ stock 2.40 ghz
18.65 on a stock 3770K which I ordered from DABS on 25/05/2012 for £234.
Series computing time : 13.32
Division time : 1.48
InvSqrt time : 0.91
Final huge multiplication time : 0.65
—————————————————————
Total computation time : 16.41 seconds(~ 0.00 hours)
CPU : i7 4790k @ Stock (4Ghz)
RAM : 8Gb DDR3
VGA : Radeon R9 380
Assuming that is pretty good for stock, things run nicely in general on my system
Total compute time of 16.39 seconds, on a 6 year old 3570k @ 4.5Ghz. Quite impressed!
Series computing time : 15.10
Division time : 1.63
InvSqrt time : 1.00
Final huge multiplication time : 0.67
—————————————————————
Total computation time : 18.46 seconds(~ 0.01 hours)
Basic 300$ build:
i3-6100
4 GB ram
256 GB SSD
Win 10 pro
15.15 Seconds.
Intel i7 4790k @ stock speed
16Gb 2.4Ghz DDR3
MSI GTX 970
250Gb Samsung 850 Pro
Windows 7 64
Stock i7-2600K
Series computing time : 17.60
Division time : 1.90
InvSqrt time : 1.17
Final huge multiplication time : 0.78
—————————————————————
Total computation time : 21.51 seconds(~ 0.01 hours)
All things considered, not bad for what is now a 7 odd year old cpu, now I'm wondering what the o/c score would be.
Edit:, now I know.
Series computing time : 14.00
Division time : 1.55
InvSqrt time : 0.95
Final huge multiplication time : 0.63
—————————————————————
Total computation time : 17.17 seconds(~ 0.00 hours)
19.31s
i7-4710MQ @2.5GHz (up to 3.7GHz)
I think that's pretty impressive considering this is a 4 year old laptop that cost less than £750.
Wish I could try out my 1700X, though seems AMD doesn't fare too well in this particular test, despite being rated as better single threaded in other tests.
ROFL…Who cares? Can I make money with that score? Can I have fun with it? What's the point?
12.88 seconds
8700k @ 5ghz using my old trusty h100i cooler from my old 2700k :)
i7 3930K@ 4.2ghz with DDR3 2400 CL 10 T1 and a 960 EVO
Series Compute time : 14.45
Division time : 1.55
InvSqrt time : 0.97
Final huge multiplication time : 0.64
Total computation time : 17.87
Who says ole' sandy bridge E is dead.
i7 3930K@ 4.2ghz with DDR3 2400 CL 10 T1 and a 960 EVO updated with screenshot.
making it a very reliable indicator of a PC's raw speed.
That really needs justification. I mean, what is it testing? I have even seen rumour that it is using 8087 FPU commands, which no sane programmer would use and are basically emulated on a modern CPU so that would be a test of how good your PC is at emulating legacy fpu code. On top of that it is a workload that no-one does on a daily basis.
Really, I switch off when I see this benchmark to the point of sometimes giving up on an article it is in. Please replace it with something good. Even Quake would be a more useful benchmark than this.
23.62 total time on stock Ryzen 1700 w/slow RAM.
Seems this one really favours Intel, will try OC'ing later and see what happens (not hopeful for a total under 20 though based on the 1800X results posted earlier).
Yay I have the slowest score!
============================================================
Total computation time : 23.11 seconds (~ 0.01 hours)
============================================================
Not bad considering a 2.6GHz Xeon E5 2670 CPU, it seems to match the Ryzen 1800x at stock.
I cant join in on this one as I run a Hackintosh
Series computing time : 13.64
Division time : 1.47
InvSqrt time : 0.90
Final huge multiplication time : 0.61
—————————————————————
Total computation time : 16.70 seconds(~ 0.00 hours)
—————————————————————
Pretty chuffed with that! i5 4670K @ 4GHz, bought off ebay 3-and-a-half years ago for £116. :mrgreen:
It seems my very own overclocked I7-7700k beats the 8700k that hexus used haha!!! 13.59!!
HEXUS - Do I get a prize for beating you on an older CPU? :)
after a few more runs the best I have seen it do is 13.50
No idea. Can get 25M digits in about 1.5 seconds on y-cruncher though. Don't really see the point of antiquated, unoptimised and totally irrelevant synthetic benchmarks myself.
watercooled
No idea. Can get 25M digits in about 1.5 seconds on y-cruncher though. Don't really see the point of antiquated, unoptimised and totally irrelevant synthetic benchmarks myself.
So 2.5 times the digits in a magnitude less time?
Guessing you aren't on a refrigerated 28 core CPU there either :D
Just a plain old 7700 at stock.
Yeah! Because Pifast MATTERS!
17.4s on a stock 4690k, which is .4s longer than
the hexus review. I'm running faster RAM than the hexus review (2.13GHz) but with looser timings
it was first 15.05 but then i closed down the browser, quit all the stuff on my 6 years ago installed windows7 on my many years old 3770K and got 14.95
So from today im calling my PC blazing fast!
I overclocked the CPU and memory about 5 years ago i even don't remember my settings around 4.7 GHz it could be (i'm getting old i know )
22.04 seconds using an i7-980 o/c @ 4GHz.
Realised I was running some BOINC stuff in the background, turned that off and managed 21.81 Woo hoo!
18.01 on an i7-6850K @ 3.7GHz (Normal BIOS settings on a ROG Strix X99 Gaming m/b)
16.7 on an i7-6850K @ 3.9GHz (Optimal BIOS settings on a ROG Strix X99 Gaming m/b)
16.41 on an i7-6850K @ 4.1GHz (Optimal BIOS settings on a ROG Strix X99 Gaming m/b, Memory @2666MHz)
watercooled
No idea. Can get 25M digits in about 1.5 seconds on y-cruncher though. Don't really see the point of antiquated, unoptimised and totally irrelevant synthetic benchmarks myself.
Right. 16.5s in PiFast to get 10m digits or 4 seconds in y-cruncher (single threaded) to get 16m.
The reason all these old Haswell/Ivy CPUs compete in PiFast is because it's ancient and doesn't use any modern extensions.
Series computing time : 17.49
Division time : 1.95
InvSqrt time : 1.19
Final huge multiplication time : 0.80
—————————————————————
Total computation time : 21.47 seconds(~ 0.01 hours)
CPU = Intel pentium g3258 @ 3.2 Ghz
edmundhonda
Right. 16.5s in PiFast to get 10m digits or 4 seconds in y-cruncher (single threaded) to get 16m.
The reason all these old Haswell/Ivy CPUs compete in PiFast is because it's ancient and doesn't use any modern extensions.
I've said it before, but it's not having legacy benchmarks (or comparing performance across CPU generations on such apps, as a matter of interest) I have an issue with, as an example Anandtech do them but clearly put them under the section of ‘legacy tests’. I
do have issue with misleading claims like implying that this single, questionable benchmark has significant relevance when it comes to ‘raw speed’ (whatever that means), putting it up front on modern CPU benchmarks and drawing dubious conclusions from its results.
If you want to know how fast these “apps and games still dragging their feet in terms of multi-core utilisation” perform - benchmark those instead! People might be interested in how Quake runs, but performance of PiFast doesn't tell us that. I doubt many are particularly concerned, when it comes to purchasing hardware, with how fast one pi-calculating application works when there are applications such as y-cruncher which will output the same result with orders of magnitude better performance on identical hardware. And neither can you use this benchmark to infer y-cruncher performance as their performance is very dissimilar, and it even includes a CPU dispatcher and different, tuned binaries for different microarchitectures and different ISA availability.
As we've seen multiple times in the past, extrapolating from benchmarks to predict either future performance or performance of other applications has major caveats and sometimes completely wrong. Plenty of review sites were criticised for running very low-res game benchmarks, deliberately causing the game performance to be CPU-limited, as a way of showing ‘how good the CPU is at game code’, which of course turned out to be wrong as game code obviously continued to evolve to take advantage of hardware.
i5 6500
Series computing time: 15.67
Division time: 1.68
InvSqrt time: 1.03
Final huge multiplication time: 0.69
Total computation time: 19.13 seconds(~0.01 hours)
I thought I'd do this as I've done some maths on the worth of a GTX 1080Ti or 1080 and in terms of improvement over my current card they're very nearly the same (in terms of percent improvement over current card per pound) but I wanted to ensure my current CPU set up (which is liquid cooled and I can't be arsed changing) wouldn't be a bottleneck. I have an i5-4690 at stock speeds (liquid cooling is there for future overclocking as it's not currently an issue but in future it may be and the enhanced cooling will improve lifespan……. okay so it looked cool, I had the money and so I did it, okay???) which scored 17.71 seconds. I thought this was a glitch but it's reproducible and others with similar CPUs seem to be getting the same. I'm going to post a thread in the forums for opinions on an upgrade, I'd be glad to hear opinions.
All right, there have been enough interesting threads and I could have enough fun with this one that I decided to register. And rather than run it on my main Sandy Bridge system, I thought why not run it on something that hasn't been done yet, like my Core 2 Extreme?
Running it on my Dell Inspiron 1520 laptop from 2007, with a Core 2 Extreme X7900 (65nm, Merom), and DDR2 667 MHz on Windows XP. Taking the best of two runs at a speed.
2.8 GHz @ 1.1625V (my usual top speed): 35.23 seconds
3 GHz @ 1.25V: 33.3 seconds
3.2 GHz @ 1.25V: 31.5 seconds
Couldn't get it to run more than a second or so at 3.4 GHz, even at the max voltage of 1.275V. But I'm pretty pleased to get less than double what some of the Devil's Canyon and Skylake samples are scoring on a 10-year-old mobile chip.
The fun part was that since this was single-threaded, I could push it farther than multithreaded workloads. For multi-core benchmarks, 3 GHz is just not possible, and at 3.2 GHz with a multithreaded workload it'd fold like a paper crane. But keep it single-threaded, and it can push out those few hundred extra megahertz.
27 seconds… Ryzen 2500u. Not bad for a 2ghz mobile chip, looks like AMD's second generation Ryzen is finally giving Intel the kicking it has needed for a while!
watercooled
I do have issue with misleading claims like implying that this single, questionable benchmark has significant relevance when it comes to ‘raw speed’ (whatever that means), putting it up front on modern CPU benchmarks and drawing dubious conclusions from its results.
The only thing to take away from the pi fast benchmark, is how quick a cpu is at running pi fast. Literally, that is it. I don't see any correlation from running pi fast, to running any other particular game or software. If I want to know how well a cpu performs running something, I'll search out specific benchmarks similar to the conditions I'll be running with (cpu / gpu and other hardware).
Is pi fast a good indicator of single threaded cpu performance? As a legacy test, perhaps. I have no idea what extensions it uses, so it may well be biased towards one particular set of cpus, which looking at this thread it most definitely is. I wouldn't think my 7 year old Intel i7-2600k is better than the AMD R7 1800X, not even remotely.
I have no idea - 3.14159 is more than accurate enough for my every day requirement, I don’t need to calculate it to any higher degree of precision.
Series computing time : 12.31
Division time : 1.31
InvSqrt time : 0.81
Final huge multiplication time : 0.54
—————————————————————
Total computation time : 15.03 seconds(~ 0.00 hours)
I5-6600K
Kinda pleased with the old girl - i3770k not overclocked, running at 3000MHz, 16Gb DDR3 at 933Mhz.
Series computing 16.39
division: 1.47
InvSqrt: 0.94
Final huge multiplication time : 0.61
Total computation time : 16.39 seconds
Duration : 15.60 seconds
i7-7700k (4.20GHz) in a Clevo laptop chassis
I notice the test gets results roughly 5 secs slower if I run the .bat with Chrome open
14.65 with an i7 3770k clocked at 5ghz
16:08 Seconds - i7 2700k @ 4.7GHz / 16GB 1600MHz RAM
Series Computing Time 13:07
Division Time 1:46
InvSqrt Time 0:90
Final Huge Multiplication Time 0.61
Total Computation Time 16:08
Still waiting for a valid reason to upgrade this CPU as it still does everything i need it to which is primarily gaming at 1440p with a GTX1070 and a bit of photo processing.
Snaga
Kinda pleased with the old girl - i3770k not overclocked, running at 3000MHz, 16Gb DDR3 at 933Mhz.
Series computing 16.39
division: 1.47
InvSqrt: 0.94
Final huge multiplication time : 0.61
Total computation time : 16.39 seconds
Doesn't look right to me, your series computing time should be less than the total computation time, and certainly not the same.
Out of memory error on my ZX81 :(
so, i7-870 at stock with RAM running at default (so presumably only 1600Mhz), and with meltdown patches running,
Series computing time : 20.34
Division time : 2.12
InvSqrt time : 1.34
Final huge multiplication time : 0.89
—————————————————————
Total computation time : 24.76 seconds(~ 0.01 hours)
Writing informations into file pi.txt
edit, and this with meltdown disabled:
Series computing time : 20.33
Division time : 2.11
InvSqrt time : 1.34
Final huge multiplication time : 0.87
—————————————————————
Total computation time : 24.71 seconds(~ 0.01 hours)
Writing informations into file pi.txt
so meltdown makes next to no difference for this benchmark
16.11 seconds
i5 4690k 3.5ghz
Pretty sure I have it overclocked, but Windows says it's at 3.5
scruffy_wuffy
18.65 on a stock 3770K which I ordered from DABS on 25/05/2012 for £234.
xin chúc một ngày mới vui vẻ và thành công …………………………………………….!!!!!
scruffy_wuffy
18.65 on a stock 3770K which I ordered from DABS on 25/05/2012 for £234.
mua may bán đắt ạ!!!!!!!!!
scruffy_wuffy
18.65 on a stock 3770K which I ordered from DABS on 25/05/2012 for £234.
UPP để bay cao hơn!!!!!!!!!!!!!!!
scruffy_wuffy
18.65 on a stock 3770K which I ordered from DABS on 25/05/2012 for £234.
UPP để bay cao hơn
xin chúc một ngày mới vui vẻ và thành công …………………………………………….!!!!!
xin cảm ơn!!!!!!!!!!!!!!!!!!!!!!!!!
Series computing time : 16.38
Division time : 1.79
InvSqrt time : 1.09
Final huge multiplication time : 0.74
—————————————————————
Total computation time : 20.09 seconds(~ 0.01 hours)
CPU - Intel Core i5-7300HQ @ 3.5Ghz
Ram - Corsair 8GB 2400Mhz SODIMM
Series computing time: 13.12
Division time: 1.40
InvSqrt time: 0.86
Final huge multiplication time 0.58
Total computation time: 16.00 Seconds
i7 4770k OC to 4.2Ghz, still pretty good I guess
Total Physical Memory Allocated = 61372 Kbytes
Computing series with index N=705226
Starting step 6
Starting step 5 …
Starting step 4 …
Starting step 3 …
Starting step 2 …
Starting step 1 …
Starting step 0 …
Series computing time : 11.90
Division time : 1.26
InvSqrt time : 0.79
Final huge multiplication time : 0.54
—————————————————————
Total computation time : 14.53 seconds(~ 0.00 hours)
Oh - forgot to say System details didn't I ;)
8700K overclocked to 4.5 GHz (playing with core ratios atm - finding a balance between noise and FPS levels) and 16GB DDR4 3200 RAM as well.
Enter FFT Size in k :
Compressed output (also useful to specify output format) ? :
Total Physical Memory Allocated = 61372 Kbytes
Computing series with index N=705226
Starting step 6
Starting step 5 …
Starting step 4 …
Starting step 3 …
Starting step 2 …
Starting step 1 …
Starting step 0 …
Series computing time : 11.89
Division time : 1.23
InvSqrt time : 0.77
Final huge multiplication time : 0.53
—————————————————————
Total computation time : 14.47 seconds(~ 0.00 hours)
Writing informations into file pi.txt
Second run with browser open as well - Weird thing is it seems to be a better time with more stuff running as well… strange lol
Cêsar (PIXON),
Series computing time : 15.99
Division time : 1.76
InvSqrt time : 1.08
Final huge multiplication time : 0.72
—————————————————————
Total computation time : 19.63 seconds(~ 0.01 hours)
Pentium G4560 | 8GB 2400MHZ
Just for fun, I tried my i5-6200U as well. It boosted to 2.7-2.8 GHz and took 24.63s
20.92 seconds
2.30 (division)
1.44 (InvSqrt)
0.95 (multi)
25.67 (total)
—-
My PC:
Intel Core i3 2100 (3.10 GHz)
8 GB DDR3
1 TB HDD
Windows 10 64-bit
Intel HD Graphics 2000
Series computing time : 21.51
Division time : 2.18
InvSqrt time : 1.31
Final huge multiplication time : 0.90
—————————————————————
Total computation time : 25.96 seconds(~ 0.01 hours)
Ryzen 1300X @3.9GHz
After looking at the few other Ryzen scores on here I decided that mine was too slow and reran the bench a few more times. Fastest result (below) knocked off 4 seconds.
That still looks a touch slow to me, considering the OC, so I may waste some more time this weekend faffing with my memory timings.
Series computing time : 18.57
Division time : 2.11
InvSqrt time : 1.28
Final huge multiplication time : 0.88
—————————————————————
Total computation time : 22.91 seconds(~ 0.01 hours)
Ryzen 1300X @3.9GHz
17.29 Seconds
Intel Core i7 6850K (running stock)
16GB Quad Channel DDR4 3000 (Geil Dragon)
20.25 Sec
Processor: Ryzen 2700X @ Stock 3.7 Ghz
Ram: 16 Gig DDR4 3200 CL14 RAM
Graphics: ROG Strix RX480 8G OC
OS Boot: WD Black 500GB NVMe PCIe Gen 3 SSD
Data Drive: WD Black 2TB Hard Drive
Total Physical Memory Allocated = 61372 Kbytes
Computing series with index N=705226
Starting step 6
Starting step 5 …
Starting step 4 …
Starting step 3 …
Starting step 2 …
Starting step 1 …
Starting step 0 …
Series computing time : 13.32
Division time : 1.39
InvSqrt time : 0.86
Final huge multiplication time : 0.58
—————————————————————
Total computation time : 16.21 seconds(~ 0.00 hours)
CPU: I7-6700K
Cores x GHHz: 4cores, 8threads @4.01
OS: win 10 64 pro (test read as win 8 pro)
GFX card: GTX 960
DRAGONFIRE_Sr
Total Physical Memory Allocated = 61372 Kbytes
Computing series with index N=705226
Starting step 6
Starting step 5 …
Starting step 4 …
Starting step 3 …
Starting step 2 …
Starting step 1 …
Starting step 0 …
Series computing time : 13.32
Division time : 1.39
InvSqrt time : 0.86
Final huge multiplication time : 0.58
—————————————————————
Total computation time : 16.21 seconds(~ 0.00 hours)
Series computing time : 12.09
Division time : 1.28
InvSqrt time : 0.79
Final huge multiplication time : 0.53
—————————————————————
Total computation time : 14.75 seconds(~ 0.00 hours)
i7-5820k @ 4.7Ghz
16.7 on an i7-6850K @ 3.9GHz (Optimal BIOS settings on a ROG Strix X99 Gaming m/b)
