Gridcoin - Computational performance chart (PFLOPS)

in #gridcoin6 years ago (edited)

In computing, the term FLOPS (Floating Point Operations Per Second) is defined as the number of floating-point equations a processor can solve in one second. It is a degree of computer performance which is useful in the field of scientific research to compare systems [1].

BOINC projects provide credits for computer calculations based on the number of FLOPS a system generates [2]. In our bi-weekly Gridcoin State-of-the-Network updates, @jringo and I include the statistics for these projects including the FLOPS per project.

In previous episodes you may have noticed that some projects were lacking this information but with the support of @GUK, I have been able to close the gap and compile the chart as presented below. The team computational performance is not in FLOPS but petaFLOPS or PFLOPS, which is 1015 FLOPS.


(Data source: BOINC project websites)

Because there are a few jumps and drops in the chart, I thought it would be useful to provide some further clarifications.


(Data source: BOINC project websites)

  • The first significant increase in computing power (1) started mid-December last year and lasted till end-January. This was the period where the Gridcoin price reached its highest level and attracted many new team members. Even after the Gridcoin price declined again, the computing power remained stable which was really positive.
  • The first substantial decrease (2) happened beginning of April when Collatz Conjecture stopped producing work-units but then started again at the end of the same month.
  • Two projects caused the second sizable decline, Einstein@home and Collatz Conjecture (3). Einstein@home stopped providing user information due to EU’s GDPR implementation while Collatz Conjecture had validation errors with their work-units.

At the end of last month, team Gridcoin reached its highest output so far (7.74 PFLOPS), which is equivalent to the 20st most powerful computer system based on www.top500.org. In the table below, you will find the breakdown of GFLOPS per project for this event.


(Data source: BOINC Project websites on 23-June)

In future Gridcoin State-of-the-Network updates, we will add the first chart to the statistics section as well and I’m interested to monitor how many PFLOPS Gridcoin is going to generate moving forward.

Gridcoin is an open source cryptocurrency (Ticker: GRC) which securely rewards volunteer computing performed upon the BOINC platform in a decentralized manner on top of proof of stake.

BOINC (Berkeley Open Infrastructure for Network Computing) is a distributed Internet platform launched beginning of 2002 and rewards participants with credits for performed work. BOINC is an application available for multiple Operating Systems and utilises the unused CPU and GPU cycles on computers to perform scientific work.

Thanks for reading. Follow me for more BOINC & Gridcoin related articles.

Image source: (1) Header, (2) Gridcoin logo, (3) BOINC logo, (4) all charts and tables by @parejan
Sort:  

Good read.

  • might be worth to separate GPU and CPU projects
  • would be interested to see team Gridcoin performance vs total project or BOINC performance
  • top three gpu-projects performance might give some false positives; what I mean is that FLOPS surprisingly is quite a muddy measurement unit and WU optimisation might have a great impact on recorded performance - if batches of WUs are less parallelized, they will be processed slower and network FLOPS will go down even in absence of hardware or participation changes
  • as per above, drawing conclusions about participation or network power changes just from FLOPS changes might lead to erroneous interpretations

Isn't it also true that Collatz only does integer operations? If so, using FLOPS for this project could be misleading.

Good point, but...
IOPS for Collatz, single precision FLOPS for Amicable, double precision FLOPS for MilkyWay... and we loose simple ONE number to several numbers, what is bad for general public. What is more, most programs use a mix of floats, integers etc. Also, at some point in Collatz computations integers might get promoted to floats, I don't know. Deeper you go, more complicated it becomes.

I think both @h202 and yourself have valid points.

It may not be perfect but at least we have a number we can use as a baseline to compare ourselves with eg. the TOP500 most powerful systems.

@hotbit @parejan

I do think most "real-world" projects (i.e. modeling physical systems) use floats. (Unless you're talking about book-keeping operations like iteration variables and counters, which I suspect only make a small contribution.) Collatz, on the other hand, exclusively runs using (integer) mod and division operations. I don't see how floats could enter those computations. More generally, I suspect the projects which substantially mix integers + floats are in the minority.

Maybe we could do an approximate conversion from IOPS to FLOPS, but not bother with single vs double precision? I do agree several numbers are bad for presentation, so definitely good to avoid that.

Thanks for your feedback @hotbit.

  • I will consider this and see how to integrate it in the State-of-the-Network updates.
  • In my weekly project updates you can monitor this in the chart "Team RAC vs Overall RAC %". I think this indicator is "fairly" accurate in terms of team Gridcoin performance vs total BOINC performance.
  • I had to think about your comments but if WU optimisations take place or if WUs can be less parallelized, doesn't this mean that the FLOPS will go up or down and the Gridcoin computational performance chart accordingly? I need to read you articles again. :-)
  • Personally see this chart as an estimation of computational performance and understand the limitations. On the other hand it does provide a trend if you look at it over a longer period.

Personally see this chart as an estimation of computational performance and understand the limitations. On the other hand it does provide a trend if you look at it over a longer period.

Exactly. FLOPS might be the the best measuring tape we have on hand at the moment.
There seems to be two forces influencing changes in the graph:

  • total hardware capabilities
  • parallelization / optimization of WUs in different batches (especially for GPU)

Thanks for periodically putting together all these statistical data! Upvoting is the least I can do as a thank you.

Thanks @hotbit, much appreciated!

Shouldn't World Community Grid's stats shutdown be included in point 3 as well? I'm sure a lot of compute going to that project was not being counted until a couple days ago...

Good question @barton26.

Although the individual stats were not available anymore, the team stats could still be downloaded from the WCG website and as such I have included them in the overall chart.

Looking at the WCG computational performance chart for the last 30 days, you will notice a decline of 0.006 PFLOPS. If you compare this with the overall Gridcoin decline of 2.27 PFLOPS from 23-june to 9-july, WCG didn't really have an impact.

Oh, I didn't realize they left the team stats up during the GDPR fiasco. That's really interesting to see how little Gridcoin's WCG participation dropped. Pretty surprising honestly.

Here is the chart from 1st of April, before it all started, where you can see that the GDPR decline is substantial in its own right but if you compare the WCG FLOPS with projects like Einstein@home and Collatz Conjecture, it is 'just' a small contributor for team Gridcoin.

Well done, that is awesome work! I think it's important that we know how our network is performing over time with that metric.

Thanks and I fully agree with you @an0n7m0us3.

By the way, thank you for asking to generate this chart on Slack because now we can include it in our State-of-the-Network updates as well. :-)

Coin Marketplace

STEEM 0.36
TRX 0.12
JST 0.039
BTC 70112.96
ETH 3549.99
USDT 1.00
SBD 4.71