valid.canardpc cleaned up lot of cpuz 1.60 & 1.61.x validations

[Christian Ney]

As you guys know CPUZ 1.60 and 1.61.x were extremely bugged and we were just beta testers, gave headaches to hwbot's moderators due to bugged validated scores and so on.

 

Now Sam from valid.canardpc "to be fare" cleaned the ranking: link

From high frequency validations done with 1.60.x / 1.61.1 / 1.61.2 (because their trust level is =0 due to highly bugged cpuz)

 

As some may have noticed already this means those scores:

Christian Ney'sMemory Clock score - 1922.9 MHz with DDR3 SDRAM at 1922.9MHz

John Lam'sMemory Clock score - 1921.8 MHz with DDR3 SDRAM at 1921.8MHz

TK-OC'sMemory Clock score - 1918.5 MHz with DDR3 SDRAM at 1918MHz

 

Have all been rejected.

 

Which brings back the Highest RAM Frequency Reached to this one (done with 1.59 and double checked by valid.canardpc guys):

Christian Ney @ Ocaholic.ch / GSkill reached 1868.3 MHz with 4096 MB of G.Skill Memory

According to valid.canardpc's ranking.

 

At hwbot many guys don't like rejected valids posted, some are good scores and some are bugs but both are rejected.

 

Now the question is, do we also have to clean hwbot from rejcted cpuz validation submissions ?

Edited August 20, 2012 by Christian Ney

[TaPaKaH]

We all know that CPUz can occasionally produce "red" validations even at stock clocks, so I don't mind having a bunch of those in the hwbot DB as long as they are legit.

[der8auer]

We all know that CPUz can occasionally produce "red" validations even at stock clocks, so I don't mind having a bunch of those in the hwbot DB as long as they are legit.

 

agreed.

 

I don't have a problem with rejected scores. The software might say "rejected" but the benchers put a lot of work in the scores so it would not be fair to just block them.

 

Ofc if the result is obviously bugged we block it.

[TaPaKaH]

does the "new and improved" CPUz still occasionally produce red validations at 100% stable clocks?

[Christian Ney]

I didn't have feedbacks saying so atm

[I.M.O.G.]

This would wipe out 3 of the top 5 frequency globals - 2nd, 3rd, and 5th. I'm sure a lot of work was put into those submissions, and they looked legit when they were made, but they were all done on versions that were badly bugged.

 

More importantly, my cpu frequency is ranked 7th globally, so I vote "yes, please move me up 3 spots". I kid, I wouldn't want to be one of the three listed above - it doesn't seem fair completely to remove the scores.

[Guest John Lam]

agree !! Christian , Are you ready for the benching ?

[Christian Ney]

agree !! Christian , Are you ready for the benching ?

 

Planned to do it live this week at Campus Party, took the rig with me in the luggage.

[knopflerbruce]

We need to decide that in each case that pops up. IMO if you have a really good CPUZ chip you also should be able to produce a 1m/pifast score that reflects performance equal to valid minus a couple of hundred MHz, and same reasoning goes for memory frequency scores. If you can't show that => we don't give benefit of doubt.

[I.M.O.G.]

I think that is the best idea yet Bruce.

[knopflerbruce]

I think that is the best idea yet Bruce.

 

Love you too:D

[Christian Ney]

Memory is tricky actually you need to have both of them (imc + mem) stable, when I clock memory very high, setup usually crash just after taking screenshot (puts both under load).

 

You can get to 1800 memory but you can show only 1600 mhz superpi so ...

Edited August 21, 2012 by Christian Ney

[knopflerbruce]

So there is a 200mhz difference between F7 and 1m?

[Christian Ney]

at least yeah, pretty much like for some CPUs, can do high CPUZ but low bench freq

[Mr.Scott]

I didn't have feedbacks saying so atm

What about socket A, is that fixed now too?

[Christian Ney]

What about socket A, is that fixed now too?

 

What was the issue ?

[Turrican]

What was the issue ?

 

since cpu-z 1.56 the validation shows double the fsb.

[Christian Ney]

ha yeah, asked to fix for Llano totally forgot for Socket A. Don't even need to debug it's the online valid, I will drop him an email, should be fixed tomorrow.

 

[Mr.Scott]

since cpu-z 1.56 the validation shows double the fsb.

Correct, and thank you.

[Mr.Scott]

ha yeah, asked to fix for Llano totally forgot for Socket A. Don't even need to debug it's the online valid, I will drop him an email, should be fixed tomorrow.

 

Keep us informed.

[Christian Ney]

Looks like it has already been fixed:

 

http://valid.canardpc.com/show_oc.php?id=2485584

[Mr.Scott]

Sweet.

[Turrican]

ok, thats nice.

[Christian Ney]

Bump

[I.M.O.G.]

Any updates on this?

 

This could affect 6 out of the top 20 CPU-z globals on hwbot, 2 out of the top 5, and the vote seems to support taking a closer look:

Christian Ney: http://hwbot.org/submission/2326765_christian_ney_cpu_frequency_fx_8150_8181.1_mhz

Name GT: http://hwbot.org/submission/2322981_namegt_cpu_frequency_fx_8350_8176.5_mhz

wytiwx: http://hwbot.org/submission/2277245_wytiwx_cpu_frequency_celeron_lga775_352_8160.94_mhz

Ananerbe: http://hwbot.org/submission/2318479_ananerbe_cpu_frequency_fx_8150_8156.76_mhz

just_nuke_em: http://hwbot.org/submission/2284616_just_nuke_em_cpu_frequency_fx_8120_8373.8_mhz

wizerty: http://hwbot.org/submission/2297410_wizerty_cpu_frequency_fx_8150_8406.34_mhz

Edited November 20, 2012 by I.M.O.G.