BeepBeep2
-
Posts
407 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Blogs
Posts posted by BeepBeep2
-
-
Nice one Woomack, especially for air/water!
-
Crazy stuff, both of you!
-
It depends if you're lucky... if someone with knowledge starts posting, it won't go that route. However, I do not see how it's obvious that one type of prime error message can only be related to one variable - in this case vcore. You need alot of data to be sure about that, and 'anyone who has ever used an AMD FX chip' won't ever have that data. You'd need to search forums to look for a lot of people saying the same thing unless you want to go on a 200 BSOD testing spree
(this is actually one of the annoying things I see in forums from time to time...).Another rant. You got questions to a macgyver-ish type of cooling idea/whatever, and the first 5 replies are all from idiots who says like 'why don't you just do it in some regular, straight forward way' when it's clear from the 1st post that it's not an option. Most forums have too many of such bunnys who cannot answer the question, but giving irrelevant advice instead
Maybe it takes a lot of data, but even on XtremeSystems, about 10-20 of us bought AMD FX CPUs on release date and started using Prime95...within 5 days, all of us knew the basic characteristics of the platform because we all noted: It ran really hot, used lots of power, and ILLEGAL_SUMOUT was common for every person with vcore or lower temperatures helping the problem. With AMD FX, if vcore is too low, Aero will start to malfunction on some CPUs and on all CPUs, threads will fail very quickly in Prime95 with "ILLEGAL_SUMOUT"...This can also happen because of heat, but in reality it's always heat/vcore/edge of stability combined. When temperatures have wiggle room then increasing vcore helps, if temperatures are 80c+ or something then it might fail regardless.
Some CPUs malfunctioned at stock too, I had one that without LLC or higher voltage would fail Prime95 at stock frequency and Aero flashing.
...
Well, lets put it like this: Any time Prime95 comes back with a rounding error, on AMD it is because of memory or CPU-NB...pretty sure that same error comes back on all platforms for memory instability, have had it on my 2600K too.
Same thing with BSODs on all platforms DDR to DDR3, IRQL, MEMORY_MANAGEMENT, PAD_POOL_HEADER, etc is almost always ram. Same type of deal.
...
For your rant:
What kind of MacGyver?
I attached a Corsair H70 to my 2600K with zip ties after inspiration from poparamiro Most people would assume you were crazy for trying such things.
The other day I was at school in my CompTIA A+ class and I was talking to my friends/other students about overclocking and hardware.
When I mentioned putting my motherboard in the oven and that it was safe to put a motherboard in the dishwasher all the kids looked at me and said "You're kidding right?..." "Or, you're just retarded?"..."Did you try it? How did that work out for you?"
-
Must be that Black Royale theme
Nice clocks NickShih!
-
-
-
CPUz work correctly?
Seems to be only CPU Frequency and RAM amount right now. Validations work though, checksum is verified as correct, so is SuperPi checksum when checked.
I suppose that's enough grounds to invalidate, since CPU-Z is missing info. If it were to work though, then what stance should HWBOT take on it? The runs are generally 20% slower anyway...
-
+Wine, not allowed?
AFAIK the rules don't state a single thing about the OS choice, I subbed two results SuperPi 1M and PiFast (Windows versions) that were running in Wine, which isn't an emulator and contains a Windows Registry just like Windows.
What about OSes like ReactOS?
I just want to know, and request the rules be amended...
-
...
...
...
....They gave him QIMONDA!!!!!!!! GDDR3 instead of GDDR5
...
What's wrong with Qimonda? 1200 MHz can be done on the 1ns GDDR3 with a little voltage
-
Not all enthusiast forums are like that. In some forums, you still can get a question answered before 10 posts are made and have the thread pretty much done.
I hope that answered your question
/thread done.
Hey thanks!
-
Hicookie did the 7.8G validation with it - previously 7.5G. Afaik, they also did some tuning on the digital PWM, but that requires a firmware flash and the tools for that are under NDA
. Still looking for a way to get that stuff without violating the NDA, though. Hicookie mentioned he got about 50-100MHz with the LN2 mode. The rest comes down to slim XP instead of big W7.[hwbot=2329722]submission[/hwbot]
Figures, The Stilt was able to get 7600 from a 7300 5800K by tuning the PWM.
AM3+ platforms wouldn't have such a problem, would they?
Nice to see 50 Mhz gain from lower idle temperatures, maybe it's dropping temps about 5-10c
EDIT:
Oh derp, he posted in this thread.
-
Please add the ASRock FM2A75M-ITX (A75 Mini-ITX) and also the A6-5400K is listed as Llano
:Thanks Turrican!
-
Full specifications: 5400K @ 4 + 4.2 Turbo, 8GB G.SKILL Sniper LV @ 1600 8-8-8-24 1T, ASRock FM2A75M-ITX and 120w Pico-PSU
-
Plz keep those for you and beier thx
Sorry, I found what I thought was a picture of you yesterday and I just want to confirm...
-
lol I've tried too much to help the situation
-
Welcome to Internetz.
Thanks
"Me gusta."
Every time, it seems the thread takes in the range of 100-200 replies when multiple issues occur that way, and no matter how many times, those who are supposed to know what they're doing and are highly regarded by the members that don't want to claim they know what they are doing, it seems like they never pick up on anything, and just grasp for straws, every time, every problem. I want to say on average 5-10 new people go there every day in that one little section to get help overclocking too.
The 160 reply thread I'm talking about could have been minimum 11-15 replies and should have been no more than, I dunno, 20 to solve every issue presented...I just don't get people, lol.
-
Okay, in no way do I want to dog on a single forum or name any names, I just want to know if this happens everywhere.
In an overclocking forum, or enthusiast forum, new members often come for help with general purpose overclocking, and there are regular members often looked at as sub-forum "experts" that always help the users and help troubleshoot their problems.
Those experts that stick in those few subforums of the website - Do they ALWAYS not know WTF they are doing, and go about the problem solving the absolute most roundabout way or seem to just f**k around and ignore very obvious behaviors of the hardware or fail to see patterns in the hardware/software errors???
Like, this one thread...this guy had a low end ASUS motherboard that turned out to be about 160 replies long before they solved the problem, and the problem won't be solved until the user gets a new motherboard installed in which case there's still chance for it to remain unsolved.
The story goes like this:
Character list:
"User", help seeker
"Power User A" <---"Resident Expert"
"Power User B" <---"Resident Expert in training"
Other misc users.
Myself, which wants to help this user with his problems.
Once upon a time, a user could not overclock past 3.X Ghz with AMD FX CPU and ASUS M5A97 motherboard.
The user was vague about his problem, but it was obvious that he had done some research already (in the same forum) and already had some ideas on what he maybe should do.
"Power Expert A" comes and says "We need these screenshots, CPU-Z tab X Y Z and HWMonitor at load to solve the issue."
"User" complies. Blatantly obvious in his "load" Prime95 screenshot is the error on 2 threads - "ILLEGAL SUMOUT"...For anyone that has used an FX CPU, or even seen other people use them, they should know that the only thing that can cause the problem is low vcore.
"Power Expert B" tells the "User" to increase the PLL Voltage. The "User" was skeptical about this - he read somewhere that he can kill his CPU this way.
"Power Expert A" interjects when the user is skeptical - he proclaims, "Power Expert B gives good advice, and sometimes that includes increasing PLL voltage."
(I'm like, yeah, okay *rolleyes* to myself while seeing this)...
"Power Expert B" then restates that PLL voltage can help, and that "Intel typically sets it too high, while AMD dropped the ball on FX Processors because you need to increase PLL Voltage for a good OC past 4.8 GHz"
(At this point I'm like OH REALLY *memeface*? I thought to myself Sandy Bridge was the one with that potential issue.)
"User" tells "Power Expert B" that he is a noob and needs a sort of "OC cookbook" so he can make his "OC".
"Power Expert B" goes, "That's exactly why I'm reluctant on the details of a 100% full OC, we gotta get you familiar with the basics first.", ...as if "Power Expert B" knows already how to fix absolutely everything that is the "User"s problem.
The day is done - "User" goes to bed.
"Power Expert B" says "It's okay, I'll be on/off all day. I run a computer store."
The next day, "User" increases PLL voltage as directed and reports back that it didn't help. (No $^#* it didn't help!)
...The "User" says it didn't help and that he couldn't even boot into Windows, freezes with a black screen. I'm going "WTF?" That didn't make it worse." I didn't realize "User" actually upped his clock from 4 GHz to 4.4 GHz this time only increasing vcore by .025v (1.3-1.32v). (...No $^#* it didn't boot!)
"Power Expert A" comes back after not posting for a string of like, I dunno, 15 replies between "User" and "Power Expert B". He says in a nutshell - "1.32v might not be enough for 4.3 GHz, and it doesn't seem like you upped PLL voltage like "Power Expert B" told you to. Also, turn all the green crap off. Oh, and it would be SO much easier if you posted screenshots of a stable run."
"Power Expert A" obviously cannot read, as the post right above his is "User" saying that PLL Voltage didn't help. By green crap, which is the kind of terminology users use there, he meant all power saving features, which the "User" was already instructed to turn off in post #2, and already turned off. This person is obviously grasping at straws.
"User" posts a stable run. Except, shortly after he realizes it wasn't stable and increases vcore on his own. At 20 minutes, one of the threads errored out. So, he increased vcore and tried again. He says he gets ILLEGAL SUMOUT at 4150 MHz+, and now windows keeps freezing at ~20 minutes and more vcore doesn't help (keeps freezing).
Nobody answers for many hours. "User" asks if anyone has any tips for him. At this point in reading, I am looking at the thread as near-laughable. Nobody has figured out still what kind of temperatures the guy has at the highest stable clock, because all his screenshots show 7 threads running or less due to ILLEGAL SUMOUT.
"Power Expert B" chimes in, saying "Run Prime95, and small FFT, because that's the best. Set RAM timings to SPD and OC with multiplier only for now. You shouldn't need more than 2.56v PLL for 4.4 GHz, above that you will need more. Do you have 1.375v vcore now? Set it. If that doesn't work, try 1.4v. It's possible your CPU is just a very power hungry chip. Try bumping CPU-NB Voltage too, have you done that yet? Temps are important too, if you're approaching 60c it's near the limit of what you can do. Oh, I keep forgetting you have an 8150 too. These take excessive jumps in voltage for overclocking at higher OC."
(I'm thinking, so many vague statements...and they apply to like, I dunno EVERY CPU AMD/intel...I was also wondering why it took this long for someone to ask what load temperatures he had, probably because everyone thought temperatures were okay from the screenshots with 6 out of 8 threads running.)
This is about the 30-40th reply.
...
...
...
...
The thread keeps going with this roundabout BS back and forth for 160 replies, 9 pages. When I finally enter the thread, after 160 posts and right after they told the user to buy a new motherboard, I tell "User" that I haven't read the thread but I give some advice on vcore and Illegal Sumout after reading the first 15 replies and seeing it was going nowhere. I then say that I'm afraid if that is the case, and it wasn't solved, a new motherboard would not help his issues.
Many members, especially the "Experts" chimed in and told me I was rude for interjecting, that I am obviously self-centered and essentially stupid for not reading the thread (because I think it's ridiculously long and all over the place) and seeing that they already covered that issue...in the 40th reply, near the end of page 3.
The forum is quite popular because of it's friendly users and good peer tech support but I find it ironic. The members of their bench team don't normally post in those subforums, but some of them (a few) share the same ways of thinking as the "Expert A/B" I mention above, mostly because some newer members to their benching team learned all they know from Expert A/B and Expert A/B really just learned from reading a guide a guy typed up in his basement 3 years ago or because they or the people they learned from assumed something to be true and they kept repeating it without experiencing that themselves and cross-checking the phenomenon with other enthusiasts.
-
ES not allowed (like in any of the competitions).
Socket is determined by the CPU. Bulldozer is AM3+, Thuban is AM3, Phenom I is AM2+ etc
Can one person sub multiple sockets?
Also, a 940BE should be considered AM2+, right?
When I checked last week, unless I was high or something (maybe?), AMD was a one sub competition, now I see otherwise
-
2400C11 Predator kit has great price/"performance" ratio but not so good ICs:(it's same as G.Skill RipjawsX F3-17000CL9D-8GBXM
What's wrong with CFR? It can be just as good as BFR depending on bin
-
Nice run!
I saw my submission dropped a place, impressive run you have there.
-
I tested this driver in 3DMark 11 and Unigine yesterday. my card was Aircooled, the boost is amazing, but card gets much hotter now... And i cant reach my clocks on air, GPU is 20 MHZ slower now, but score 15% Higher, lol... Seems like AMD is buying performance with current
Just lower the GPU clock then
-
-
-
Linux/Mac OS
in Benchmark software
Posted
It should be written then!
Both those OSes with Wine run the bench slower than native Windows...of course, all the DLLs are rewritten but you could easily install all Windows 7 DLLs too or something.
I guess we are worried of someone finding a way to cheat or using a native Linux version of something?