yosarianilives

Well for one it doesn't have speedstep, it also doesn't support hyperthreading, it has a stock tdp of 35w, 1mb of cache, comes at a stock 1.6 ghz, supports ddr3 1066, and has intel hd graphics

Actually was gonna suggest G440, it has several advantages that are not immediately apparent.

No spoilers but you're missing a few semi-obvious bombshells

Can't wait until you fill your dewar! Did you test for max fsb?

What constitutes a Vega 56 Nano? Is it only the powercolor vega 56 nano? Or would cards like the saphire vega 56 pulse or XFX Vega 56 Double which use the nano PCB also count as Vega 56 Nano since there's nothing besides the pcb that differentiates a nano? They don't even show up different in OS or have a specially marked die. As far as I can tell Vega 56 Nano is a duplicate listing for Vega 56 since it's just a different PCB made by some manufacturers and is still a normal Vega 56 with no differences in the die or memory. It's as much a different vega 56 as the liquid edition vega 64 is a different vega 64.

What's the cold bug on this chip?

Did you find that mem OC helped score on the latest opencl drivers with AVX?

It's okay, he'll learn. We all learn after a while that we're just a very small fish in a very big pond in this hobby, sometimes even after getting a global gold. I know I still have a long way to go before I can feel like anything is owed to me by some community, probably will never happen.

bdie is happiest 5-15c, below that it starts getting angry and won't do good mhz and timings. Although for max mhz runs you let timings out and run bdie on ln2.

Get a gigabyte? lul Not sure why anyone would do that to them self, would have to be a proper masochist as every single gigabyte board I've ever worked with has been just absolutely terrible. Maybe if they knew how to make a bios that worked, or maybe if they could have said bios actually properly apply the settings it's given, or maybe if their qc existed at all. Or maybe if they just let someone who knows how to make motherboards make a motherboard then slap the gigabyte branding on it, then maybe I'd consider a gigabyte board. But otherwise I think that it's just not worth the pain for zero gain. Anyways from my experience the evga dark is very good for skyx, if not possibly the best. However it doesn't work very well for kabyx so if you want to use 7740x then you'll want an OCF. The rampage board from asus is also pretty good for kabyx but doesn't have the vrm to stabilize clocks on the bigger skyx chips.

I like this idea, but maybe add an option to pay $40 for a lootbox with a random global gold inside?

Which set of holes did you try? For lga 115x it shouldn't be the same hole in every corner.

Based on the fact that a locked cpu at 4 ghz is scoring like an 8700k at 5.2 ghz I'm gonna go out on a limb and say that this score may be glitched or he knows something about efficiency that nobody else does.

Oof, double cpu tab strikes again, thanks for letting me know early! Will resub a better score when I get the chance.

Just spitballing here, but maybe you can pay for a "boosted" period of time on hwbot. Like pay $20 to have all scores that you submit within that timeframe have their points multiplied by 2x before they're applied to your ranking, or $40 for 3x? Seems to work well for some others.

Only board I killed with water was salt water and that one was actively arcing while I was benching... I would say definitely a good idea to clean boards out after use. A good iso flush or spraying all sockets out with electrical contact cleaner works well.

This has been mentioned I think, but why is it even allowed to sub a dual gpu score for timespy stage? https://hwbot.org/submission/3982468 of course this score would be removed for lack of background anyways...

It's an awkward cutoff for points, no doubt caused by relatively few scores being within the cutoff but definitely lines up with how scoring is supposed to work.

Ranking definitely looks weird, the "70% slower than 1st place" cutoff should be around 15s 258ms by my napkin math

I've had puddles of water on completely uninsulated boards, water from the air isn't as good a conductor as you'd think. Things just start getting fucky when you get it into mem slots etc

I would think anything memory related would be heavily board dependant. There are a number of memory timings that are set by bios that you don't have access to set so certain boards will just perform better on platforms like AM3 where relatively few timings are exposed to the user. That's why for example the Crosshare IV Extreme has been shown to have better "IPC" than the Crosshair IV Formula in 32m and probably anything else that memory heavily impacts.

Yeah I personally wouldn't have the balls to cut on a l3014 myself, cutting the plastic part of the socket is much lower risk and not nearly as scary as some other things I've done (mostly cutting the back off of an pcie x4 slot to run an x16 card.)

r15 definitely gains from mem on these chips, just a small gain. Got to remember that at 6.7 on ddr2 it's just a few points behind l3014 on ddr3, and at 6.1 on ddr3 the 150 is about 10 points away from l3014. Definitely worth binning 890fx boards to find one that can support 6.5+ on a sempron.

Really interesting chip, can't wait to see more results with it/get my hands on one.

I would like to report a bug in the configuration for the gpupi 1b single core comp limitations @Leeghoofd Example A: https://hwbot.org/submission/3978772 Fastest damn single core I've ever seen