Weird rankings (wrong submissions order and points)

[havli]

Ran into these two:

 

http://hwbot.org/benchmark/cinebench_-_2003/rankings?start=0&hardwareTypeId=processor_2289&cores=4#start=0#interval=20

 

http://hwbot.org/benchmark/cinebench_-_r15/rankings?hardwareTypeId=processor_5521&cores=6#start=0#interval=20

 

Both points and positions are wrong it seems.

Perhaps there are more than just two, I'll post more links if I can find something else.

Edited February 5, 2018 by havli

[havli]

Broken points here as well http://hwbot.org/benchmark/3dmark2001_se/rankings?hardwareTypeId=videocard_1103&cores=2#start=0#interval=20

[havli]

Broken points here https://hwbot.org/benchmark/cinebench_-_2003/rankings?start=0&hardwareTypeId=processor_4260&cores=2#start=0#interval=20

and most likely here (17.5 points for 1st place out of 3 doesn't sound right) https://hwbot.org/benchmark/cinebench_-_2003/rankings?start=0&hardwareTypeId=processor_857&cores=1#start=0#interval=20

[yosarianilives]

Just now, havli said:

Broken points here https://hwbot.org/benchmark/cinebench_-_2003/rankings?start=0&hardwareTypeId=processor_4260&cores=2#start=0#interval=20

and most likely here (17.5 points for 1st place out of 3 doesn't sound right) https://hwbot.org/benchmark/cinebench_-_2003/rankings?start=0&hardwareTypeId=processor_857&cores=1#start=0#interval=20

I think rich has said that the current cause is too many people are hitting recalc at the same time as a change was pushed. So if people will be patient and not hit recalc, or possibly even sub, then it should sort itself out in time. Similar to the initial roll out of rev 7.

[ADVenturePO]

HWBOT should definitely ask  for no recalculation pushing for 2 days at the main page than.
 

 

[yosarianilives]

6 minutes ago, ADVenturePO said:

HWBOT should definitely ask  for no recalculation pushing for 2 days at the main page than.
 

 

Ideally they'd be able to temporily disable that feature/disable new subs. At the very least it may be a good idea to just take the site down for a day or so to let the server catch up. Hopefully once we get all the major kinks worked out of rev 7 we don't see as much situations like this.

[bolc]

5 hours ago, yosarianilives said:

I think rich has said that the current cause is too many people are hitting recalc at the same time as a change was pushed. So if people will be patient and not hit recalc, or possibly even sub, then it should sort itself out in time. Similar to the initial roll out of rev 7.

hey dude
which change are you referring to ? some global score ranking change was implemented?

[yosarianilives]

1 hour ago, bolc said:

hey dude
which change are you referring to ? some global score ranking change was implemented?

Basically there was a change that was pushed to production to help with certain subs not getting ranked/pts and I guess it required a recalc of a bunch of rankings. So when you get such a massive recalc and people keep hitting recalc on top of it then the server has a hard time with all the requests.

[bolc]

do you mean rev 7 recalc went through but was incomplete, and a whole recalculation had to be done again? 
hopefully the system is smart, when recal is pushed by a user many times it only does it once ...

PS: i am still jealous for the  3470  how many did you bin? I need to test this M5E check if it beats my M5F...

Edited March 16, 2018 by bolc

[yosarianilives]

44 minutes ago, bolc said:

do you mean rev 7 recalc went through but was incomplete, and a whole recalculation had to be done again? 
hopefully the system is smart, when recal is pushed by a user many times it only does it once ...

PS: i am still jealous for the  3470  how many did you bin? I need to test this M5E check if it beats my M5F...

Would you believe that was one of the chips I haven't binned? Didn't even buy to bench specifically, originally went into a machine for one of my friends, then I got it back when I upgraded him to my crap 1500x. Was happy about the imc though, when I tested it on my ocf a few months ago it seemed to get worse the higher the bclk and wouldn't break 2000 really. Now it does 2400 for me and gets some very nice results. Much better experience after all the ivy pentiums that my mve murdered in cold blood trying to train unstable mem settings in ln2 mode.