yosarianilives

Is this like a reverse sandbag? Drop a monster score early so other teams just give up on that stage

What I'd like to know is why do semprons have so much IPC, even higher than l3014 http://hwbot.org/submission/3133277_aleslammer_black_hole_benchmark_sempron_130_7990_marks

Pretty sure that mine can do 2300+ but I'll be running at 109.9 for gpupi where I have to drop to the 1866 divider so we'll see how that scores with more core lower mem. Worked for x265 and XTU for me.

I'll do a run before I do my gpupi run and see how it compares. Could be it favors memory bandwidth over core speed when you run 4 threads on a 2 thread cpu.

Not sure about others, but I have an l3014 that I haven't had time to mess with because of other projects and competitions. So this stage is a good opportunity for those who haven't had a chance to bench their chips to go for some 1c globals. They're not rare, just pricey for fake internet pts (until you consider the cost of literally any other globals excluding XTU and X265)

I think without the L3014 that stage wouldn't be. But I agree that all server hw that has similar specs to desktop chips (intel pre lga 2011, AMD s939, etc) should be allowed.

If you're referring to them being slanted that's how all R6A boards are

Agreed, TSE uses 6+ gb of vram. Trying to run it on a 2 gb gpu is gonna be rng.

Why have I just ordered a gt 420 on ebay now? (hard to find a pci card)

qx9775 is consumer cpu on a server socket. Will leave it up to leeg to make the final determination.

gpupi 3.2 works just fine on this install for context

Hi! I'm having trouble getting 3.3 to run on my G470, I've tried a few different batch sizes as well as hwinfo detection levels and it still crashes on run. Attached is a screenshot and error logs, I can provide any other info that you need. @_mat_ HWiNFO64.DBG GPUPI.cfg

Why wouldn't it be? Reference frequency is a score for a motherboard. So socket would be determined by the motherboard used. Pretty sure it's always worked that way.

I can only assume it's to keep people from benching until tomorrow. Otherwise I know I'd already have something pulled down...

Bump, gonna give anyone who didn't see this yet a chance to vote then we'll start planning the meet/deciding if there'll be one.

Also don't think it ever got clarified what sockets are combined. Obviously 1151 is a single socket, but what about am3 and am3+ or am2 and am2+? Will they get combined?

Personally I would prefer not to see titans however I also would be fine with whatever decision is made.

@Leeghoofd Rules don't seem to specify that titan's are banned. So then can we assume that titan cards are allowed for all 3d stages?

Which is funny cause Vega M is actually polaris based while the APUs are Vega based...

If there's one thing I learned last year, it's that it wouldn't be a country cup without at least one benchmark that takes a minimum of 30 minutes per run and needs cold.

Not if it's divided by socket, and my understanding is that all 1151 is considered the same socket. So you could have 2x modern intel scores.

This is a fairly low priority but I've noticed that some results for android benches like pcmark are done on x86 pcs running an android emulator like remix OS. Is this allowed or not?

So less than 24 hrs until comp starts and still the comp page is a bit bare. Is it gonna be completed with all stages and limitations by start of comp? I ask because I have still half a dewar left from TC and want to get a running start over the weekend if possible. Would be nice to know what exactly I have to prep tonight.

I think that being able to join a team specifically for the comp is a great thing for those on smaller teams. Coming from a team that had a few people do just that I see nothing wrong with it. They definitely contributed and helped us achieve second but I don't feel that they're the reason for our ranking. If they had been the only ones to sub we would've ranked much lower. It was a combined effort of the whole team. I'm also for allowing new members to join a team mid comp, there's no reason not to allow this, and not just because it helps our team a lot. (definitely recruited a 2x titan x pascal system from reddit with 36 hours to go on the comp, among other hw nobody wanted to buy throughout the comp) I think that encouraging participation in comps in any way as long as validation rules and such are followed is always a good thing and will only help to grow the community.

correct, spd is only required for the mem freq stage