MOA 2013 Discussion

[IanCutress]

Then isn't it all arbitrary then to who sets the 100% score (or wherever it comes from)?

If we put FS stage to 10000000 = 100%, then 1000pts is 0.01% and positions change.

In sport it's all about who finishes the line first over a number of rounds and gets points, not who had the fastest overall time across each track.

 

/rant

 

Edit: I approve of the front page scoreboard. Need it for final few days of ProOC.

[Splave]

I believe the percentage is based on how much faster they are then a "target" score so the number is not arbitrary nor set by any of the competitors. I like your Idea of format better, but leads to the great chance of a tie.

[IanCutress]

I believe the percentage is based on how much faster they are then a "target" score so the number is not arbitrary nor set by any of the competitors. I like your Idea of format better, but leads to the great chance of a tie.

 

The target score is set by someone (either picked or benched), thus is therefore by definition 'arbitrary', even if it relates to a score someone got. It means in the end that more time+effort+skill into bench A means more of a % rise over time+effort+skill into bench B.

 

I don't know why, but in general US sports seems to be afraid of a tie.

Elsewhere in the world, it's regular to get a tie.

But just because you get a tie in a round doesn't mean at the end people will tie.

If two people end on the same points, it comes down to who had the most 1st places.

If that's the same, then it comes down to who had the most second places.

And so on.

This isn't a radical idea, or something I randomly thought of. It's common in racing and other sports (link: rule 7.2, link, link).

 

Just sayin'.

[Splave]

lol not afraid of tie man, 1st 2nd 3rd VS 2nd 3rd 1st would require a tie breaker and Im sure all the nice MSI employees want to stick around for a couple more hours to watch

[IanCutress]

Nah, that wouldn't need a tie breaker. The first person beat the second in two out of the three benchmarks.

[I.nfraR.ed]

Although I'm not competitive at the highest level, I can only agree the format is a little unfair.

IMO this is the biggest flaw of MOA.

Ofcourse not saying T0lsty is not a great overclocker, sure he is.

Congrats for the win!

[der8auer]

I believe the percentage is based on how much faster they are then a "target" score so the number is not arbitrary nor set by any of the competitors. I like your Idea of format better, but leads to the great chance of a tie.

 

Yes there were baseline scores done by MSI:

 

SuperPI 32M: 7min 03.441 sec

Cinebench R11.5: 8,93 points

3DMark Fire Strike: 9891 points

 

And then keep in mind that 32m and FS counted 40% each to the final score and Cinebench 20%.

 

This way it does not matter if you beat someone else in one of the single benchmarks. However you have to have very good and high scores in all three benchmarks to win.

 

I really like this concept. Makes it a lot more competitive at the end than any other concept.

[IanCutress]

This way it does not matter if you beat someone else in one of the single benchmarks. However you have to have very good and high scores in all three benchmarks to win.

 

But a 1st+1st+6th lost to 5th+9th+1st. Any way you slice it, that's just a little odd. If Alva had dived the last benchmark and had 10th then I'd understand, and as Cine was only 20% you could argue that the second round was more 1st vs 4.5th.

 

I appreciate that the way it was run is more compelling viewing (harder to predict, etc), and pulling nothing away from T0lsty because he had to show the skill and push the gear to get where he was. Anyone looking at the positioning, or new people to the scene not understanding how it works, is going to think 'wtf'. I know I did - I only caught the tail end of the results.

 

Kind of odd they didn't select:

 

Spi 32m: 7min

Cine: 9.00

3D FS: 10000

 

Why not round numbers? I bet if you stuck these numbers in, someone would swap a place.

 

In b4 'these are the rules, either work with or work without'. That's true, no harm in discussing/debating the merits, even if pointless/no impact

[der8auer]

I think those numbers were just done with one of the setup at default settings/speed which makes sense if you want to do Overclocking in percentage.

[IanCutress]

I think those numbers were just done with one of the setup at default settings/speed which makes sense if you want to do Overclocking in percentage.

 

Yes there were baseline scores done by MSI:

 

SuperPI 32M: 7min 03.441 sec

Cinebench R11.5: 8,93 points

3DMark Fire Strike: 9891 points

 

And then keep in mind that 32m and FS counted 40% each to the final score and Cinebench 20%.

 

This doesn't make sense looking at the numbers.

 

Final score for SniperOz = 54.34+28.65+58.25 = 141.24, meaning 100% on each

 

SniperOz's scores:

Spi : 311.703 seconds, +54.34%, meaning 311.703/(1-0.5434) = 682.66 seconds original calculating back

Cinebench: 12.79 points, +28.65%, meaning 12.79/1.2865 = 9.94 points original calculating back

3DMFS: 14404 marks, +58.25%, meaning 14404/1.5825 = 9102 points original calculating back

 

Final score for Zzolio = 53.92+29.25+54.6 = 137.77, meaning 100% on each

 

Zzolio's scores:

Spi: 314.141 seconds, +53.92%, meaning 314.141/(1-0.5392) = 681.73 seconds original calculating back

Cinebench: 13.06 points, +29.25%, meaning 13.06/(1.2925) = 10.10 points calculating back

3DMFS: 13500 marks, +54.60%, meaning 13500/(1.5460) = 8732 points original calculating back

 

There's something wrong with the maths - people are being compared to different scores and the % are all wrong. They're not even near the numbers you quoted. It follows all the way down the leaderboard - people are being compared against different numbers.

 

More than welcome for someone to correct my maths if they think I'm wrong.

[TaPaKaH]

SniperOZ's 32M of 311.703s is a 135.85% factor of the reference 423.441s. Multiply that by 0.4 (weight of the benchmark) and you get the 54.34%.

% sign might indeed be confusing, think of it as points: in 32M you get 40 points for matching the ref score and 0.4 points for every percent, for which you improve it.

Edited October 18, 2013 by TaPaKaH

[IanCutress]

The mathematics works, but there are two issues:

 

1) The % mark is a bit misleading and not technically accurate.

 

It's not a unit that carries on after multiplication of a scalar quantity.

+54.34% means nothing in that regard, as it's not +54.34 percent of anything.

54.34 points would make more sense.

 

2) Super Pi:

 

SniperOz = 423.441 / 311.703 = 135.85%

Zzolio = 423.441 / 314.141 = 134.79%

 

You're changing the denominator, meaning that moving from 315 seconds to 314 seconds has less effect than 310 to 309:

 

e.g. 423.441 / 315 = 134.425%

e.g. 423.441 / 314 = 134.854%

difference = 0.429%

0.429% of what? you've changed the denominator

 

e.g. 423.441 / 310 = 136.594%

e.g. 423.441 / 309 = 137.036%

difference = 0.442%

0.442% of what? you've changed the denominator

 

It works when counting up, line in Cinebench:

 

e.g. 10.00 / 8.93 = 111.982%

e.g. 10.10 / 8.93 = 113.102%

difference = 1.12%

1.12% of what? 1.12% of 8.93

 

e.g. 11.00 / 8.93 = 123.180%

e.g. 11.10 / 8.93 = 124.300%

difference = 1.12%

1.12% of what? 1.12% of 8.93

 

Bottom line: The way of calculating the Spi results means the faster your result, the exponentially better your % is.

 

Next time perhaps. Or maybe it's better that the one who is in the lead deserves earning more overall points for their 1st. (Another reason the points system in F1 has changed over the past decade, so 1st places are more significant.)

[Xtreme Addict]

Firstly I really want to congratulate T0lsty! His GPU was awesome and he did very well pushing it to the limit! 2x times 1st place during 2 weeks - AWESOME!

 

Also of course Lucky_n00b also performed great scoring 2nd place! It was a tight battle in the end! Congrats Alva!

 

With % points on MOA it's always ALL about GPU in fact

 

Event was VERY good this year. Benching place was nice. I was suprised that MSI decided for a closed event, cause it's always better for marketing, but for us ocers, this place was very good. We had a lot of light (not like last year - dark, loud, a lot of viewers which don't help during benching ). Only in my opinion preparation time was a bit too short, I had to use a lot of time from SuperPI to end system mounting, especially that it takes much more time when you are benching alone.

 

In my case: I got good cpu, but not good gpu (very weak core).

 

SuperPI 32M - I failed this stage, I had a lot of problems with memories, couldn't set them right and achieve high frequency, I had very bad efficiency with RTL57 and so on, but I couldn't tighten it, finally in the end I managed to tighten a bit main timings, but I had time only for one run and efficiency was overall very bad for those clocks of cpu. I am pretty sure that my memories were dying cause after SuperPI run I went to BIOS to save profile and it freezed and both memory sticks were dead... Good that MSI/Corsair prepared for everyone replacement!

 

Cinebench - it went okay, I had a bit higher score but it crashed during screenshot.

 

3DMark - my GPU was not good 1730 GPU to achieve was a real fight. Memories were okay (1900 but no more). My GPU couldn't bench with -90/-100 caused 3D crashed. I had to bench at -80*C.

 

This contest for me went smoothly (comparing to AOOC 2013 - in Moscow we had a lot of strange issues, problems - simply with Perica we had bad luck like dead cpu, problems with software, drivers, motherboard and so on). This time I had no problems with motherboard, cpu and gpu itself, everything was working okay, only on first kit of memories I had problems, but second was working on air better than previous on ln2 I remounted platform completely 2x times, first time after SuperPI (memories appeared to be dead, I thought that motherboard simply froze too much), and second time after cinebench cause I started to have some isolation issues (wet dram slots) so cause I had time, I warm up everything and did new isolation. But MSI hardware worked very good, stable, Mpower Max and Lightning are strong pieces of hardware, only I had bad luck with bad sample of GPU (My elpida card is flying 1800+ easy but Elpida efficiency...). If both GPUs were on the same memories, I think that final ranking could differ a lot

 

All in all I am very happy. Top 3 on such worldwide event competing against best ocers in the world is a very good achievement. This MOA Final was organized very well. Now looking for more fun during Freestyle battle.

Edited October 18, 2013 by Xtreme Addict

[dinos22]

Well done guys, that was an awesome event. i really enjoyed watching you guys go at it.

 

T0lsty is the man. He had a poor clocking CPU but it didnt stop him, he minimised the damaged in first two rounds and then smashed it out of the park in 3D. Vivi was the only guy that was more efficienct running his RAM subzero but it went pair shaped for the young fella which is a shame as it made it hard to compete for later rounds. He would have been looking tough to beat had the rig been smooth sailing but what can you do! Well done to Alva and Michal. Alva's last run before the final gong just got him into 2nd place. Nerve wracking and we were there to witness it all as he was having his rig broadcast on livestream for us to see! Awesome!

 

Well done to OC TV guys. This format is really good. Shame about YouTube pulling the broadcast which resulted in some problems but all is well that ends well right! I liked the multi monitor format and you got commentary going in the end. That made it a lot more fun to watch for us! Thanks!

 

Proud of my TeamAU boys. They had pretty bad CPUs but they kept at it and cranked it pretty hard in the last round with 2nd and 3rd places in 3D MARK. Michal mentioned that this comp is all about 3D and it is. Nothing wrong with that. It requires a lot more skill to put it all together as you're benching the GPU and CPU. Surprised that more people didn't freeze the RAM as Vivi did. It clearly worked well for him in first round. That would have made it pretty tough in a live comp i guess though and certainly a lot more complicated once things go bad. Anyway, awesome work TeamAU boys! Hopefully better luck next year and expecting to see that monster setup from Carl today singing nicely

Edited October 18, 2013 by dinos22

[Lucky_n00b]

Morning folks,

 

First of all, Congratulation to T0lsty as MOA 2013 Winner! Very efficient run and damage control on the 2D benchmarks, and KILLING it on the 3DMark Fire Strike - EPIC! It was a intense battle from the start to the very end, my buddy Michal did his best with his setup and battling many problems, and finally managed to get the 3rd place - Congrats bro!

 

Also good fight from the Aussies, Jack and Carl, who pushes it to the limit with their 3D runs. Had they got a better CPU, top 3 wouldn't be impossible .

 

Vivi pushes a very efficient run in Pi 32M with the cold RAM setup (the most efficient run IIRC, with T0lsty being 2nd in 32M efficiency), but too bad the poor clocking CPU and the motherboard give him many problems, better luck next time man!

 

As for me,

From the beginning I wasn't expected a top 3 finish, let alone placing on the runner up among with the best overclockers in the world. I'm happy enough to somehow 'slip in' through the qualifier

I'm just really lucky having a good CPU (my 1st lucky CPU after 5 Years of MOA, yay! ). The CPU clocks quite well, in 4C/8T config it's inferior to Michal's 6.2Ghz Cinebench, but does quite nicely in 2-Core.

That being said, I'm still a bit annoyed looking at my Pi 32M RAM setup, didn't manage to get the 2800 and 2933 Multiplier working, so had to change to 125Mhz BCLK and clocked the RAM at DDR3-2750Mhz. Pulling it down with 10-13-12 timing didn't help much for the efficiency, only 5m 15s-ish without CW in the 6Ghz mark. I prepped the system to run RAM on LN2, but didn't manage to do it since I was troubleshooting so many times in the Pi 32M setup.

 

For the GPU, I was struggling with the Elpida-equipped GPU, the Core was better than my Samsung one, but the Mem was total crap (3300Mhz-ish). Had to change to the Samsung Card literally in the last 50 mins of the competition, very nerve wrecking to pull out a good score from a Card which I haven't tested at all before in the next 40 mins.

You guys watching my live desktop seen how many crashes I got after changing to my 2nd GPU, it was horrible X_X. The funny thing is, with the very intense competition going, I didn't realize that my live desktop was still on . In the end, the scores came out in the very last minute - again, lucky .

 

Very good event from MSI and great coverage by Hwbot & OCTV! Keep pushing it guys

 

Cheers,

[Massman]

Kind of odd they didn't select:

 

Spi 32m: 7min

Cine: 9.00

3D FS: 10000

 

Why not round numbers? I bet if you stuck these numbers in, someone would swap a place.

 

In b4 'these are the rules, either work with or work without'. That's true, no harm in discussing/debating the merits, even if pointless/no impact

 

The scores were selected based on the expected best-case-scenario score, which is the WR for each platform (so Haswell). From that WR, we took something like 70% to set the baseline.

 

It used to be round numbers, and that is arbitrary. By using a reference score to get the baseline of, there is some logic behind it. If you take arbitrary numbers, there is a benchmark scaling bias. By this I mean: with arbitrary numbers, you don't take into accoun the scaling characteristics of the benchmark. You end up with the situation where the SuperPI-32M score has almost no effect in the final score.

[Xyala]

And here is the OverClocking-TV playlist of this MOA 2013.

 

Enjoy, and as usual: Feel free to share, re-post, re-use etc.

 

http://bit.ly/msi-moa-2013

[Xyala]

Also, a big thanks from the OCTV crew for having attended and watched the livestream. In order to pitch our ideas, improve things and keep pushing it, we would appreciate a few minutes of your time to feel the following survey:

 

http://bit.ly/live-feedback

[ObscureParadox_3XS]

Had a blast watching this and chatting with a few of you guys in the chat section, even if a few of us went a little too far sometimes

 

Country cup now but roll on next year!!!

[Xyala]

OC freestyle day results , New world record!!!

 

English: http://www.overclocking-tv.com/content/events/19190/msi-moa-2013-freestyle-competition/

French: http://www.overclocking-tv.com/fr/content/evenements/19193/msi-moa-2013-competition-libre/

[rbuass]

Thanks HWBOT, MSI and all involved in this great event.

That's amazing can meet the elite of overclocking... and I am happy to can be there.

My personal view... is:

1st day

I had no lucky... got a Samsung card worst than my old Elpida... maybe the worst VGA ... and a -120 CB crap 5700 Mhz CPU (Spi)... but regardless all... I was happy...

Tolsty is really a great Overclocker and earned... Lucky Noob and Xtreme Addict also made a really good job... and surprised with the new guys (JJJC, Mike, FTW and so on)...

This happens... even the overclocker is a "master top skill"... can have no lucky... this happens, for example to Cyclone, that is really good and were not in his lucky day...

What I mean is that is not enough to be very good, but need to be in a lucky day...

So... I only have to cheer to the champions and all.

 

2nd day

I focused, since I leaved Brazil, Unigine Heaven, ...

The reason is because I prefer 3D... and also see no options, since I have no a second Titan... and before changes, the World Record was overall... so I give up to try a GTX 780 beat Titans in 3DM11 and Fire Strike.

My goal was to dispute with a master Sniper OZ... that sent a greatttt Vantage score few days ago... warming to MOA.

The first minutes, my card could not work...

Gnidaol saw me too nervous when I found a damage hardmodding, for a GPU voltage.... in the middle of many wires insulated...

ATM I was lucky to found the problem and fix there...

So... the day run smooth to me and I could bench for 8 hours without any issues.

I feel happy to got my acomplishment...

 

For the last, I want to tell that ALL the guys was there are really good... and we had very good fun for this days.

One more time, MOA was unforgetable.

 

Best wishes for all

Edited October 20, 2013 by rbuass

[Booj]

Well done everyone.. great competition this year on the first day with everything coming down to the last submits.

 

Proud of our Aussie boys JJJC and SniperOZ to finish where they did despite weaker CPUs.

 

Crazy effort by T0lsty, Rbuass, Lucky N00b, Vivi and everyone.. bring on MOA 2014!

[ObscureParadox_3XS]

Just a thought with future competitions. Would it not be better and make it more about what the competitors are able to do in terms of efficiency if they were all given kits with the same batch numbers. Then at least there is more of a chance of their being similar overclocks between them all? Or am I just talking out my arse and should shut up now??

[Mikecdm]

This was my first time at an event like this and there is really nothing that you can do to prepare for it if you haven't ever competed live. I practiced at home and everything was solid but once everything went live, nothing plays out as you had planned. The winners were pretty much those who were able to make the best out of the situation. Almost everyone that tested the hardware at home said it was solid, but at the event we were all having a difficult time running the ram correctly. My cpu was faster than several yet I probably had one of the worst times. I was planning on freezing the ram, but the way things were going, I figured it'd be safer to just run them on air.

 

The experience was amazing though. It was nice to meet everyone and the MSI staff was awesome. Congrats to the winners, one can really appreciate the effort and hard work that it takes to win at an event like this and for T0lsty to do it at back to back event, thats just incredible.

[Massman]

Just a thought with future competitions. Would it not be better and make it more about what the competitors are able to do in terms of efficiency if they were all given kits with the same batch numbers. Then at least there is more of a chance of their being similar overclocks between them all? Or am I just talking out my arse and should shut up now??

 

This might be an interesting read for you: The paradox of a fair overclocking competition