Overclock.net banner

621 - 640 of 5072 Posts

·
Tinkerer
Joined
·
196 Posts
I want to get 32GB running, is 3200MHz CL12 feasible? (4x8GB 4000MHz Samsung B-Die)

Would it be easier to obtain (if at all) with 2x16GB DR or 4x8GB SR?
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #623
Quote:
Originally Posted by RossiOCUK View Post

I want to get 32GB running, is 3200MHz CL12 feasible? (4x8GB 4000MHz Samsung B-Die)

Would it be easier to obtain (if at all) with 2x16GB DR or 4x8GB SR?
Hi, CAS12 regardless of the high bin will likely need 1.5v or in excess of. This goes for both - however, you will have an easier time with the 8GB modules.
 

·
Tinkerer
Joined
·
196 Posts
Quote:
Originally Posted by Silent Scone View Post

Hi, CAS12 regardless of the high bin will likely need 1.5v or in excess of. This goes for both - however, you will have an easier time with the 8GB modules.
Happy to go over 1.5v, might give it a whack
smile.gif


Figured as much with the 8GB DIMMS, which is why they're already on order.
 

·
Meddling user
Joined
·
7,423 Posts
The Stilt used greater than or equal to 1.45V for C12 tight, he also had Geardown on, on F4-3600C15D-16GTZ. OP of my thread Is RAM MHz king? has info.
 

·
Registered
Joined
·
4,200 Posts
Quote:
Originally Posted by Robenger View Post

How are you running memory that fast on a B350 mobo?
you think thats fast?

look at the whole ram table thing
biggrin.gif



@gupsterg
thats what im working on right now with those ram timings. drop it down to 39.50x did 2 hrs on Y-Cruncher decide is time for a prime blend

40x seems impossible to have stable for this 1800X/MoBo combo with those rams tweaks, maybe 3200 can do it.. But the faster the ram the more stability in question.
But you know we have people calling this setups stable without proper testings..

yeah they are very bencheable at everything even @ 41x but its not something i call 100% stable, i can get away with it casual gaming and regular computer use they are RB/x264/gsat/hci stable etc. Something that i call stable 100% no.. Specially knowing and what apps to use and specially where to hit it lol.

See according to my testings

here
http://www.overclock.net/t/1624139/official-ryzen-7-1800x-1700x-1700-owners-club-4ghz-club/16190#post_26279104
Quote:
So sum it up, i didnt notice no difference in this particular game on 3200 16CL vs 14CL

4.1GHz

2133 to 3200 19%+
2133 to 3466LL 14/15/15/15/28 51%+
3200 to 3466LL 27%+

3.8GHz

2133 to 3466LL 14/15/15/15/28 41%+

3.8GHz vs 4.1GHz with 3466 LL its only a 8%+
If i have to loose a multiplier i will i get more gains out of ram tweaks than mhz
 

·
Registered
Joined
·
330 Posts
Quote:
Originally Posted by zGunBLADEz View Post

C:\Windows\System32\bash.exe
So when i type in "stressapptest -W -s 3600" I get the following:
Log: Commandline - stressapptest -W -s 3600
Stats: SAT revision 1.0.6_autoconf, 64 bit binary
Log: buildd @ kapok on Wed Jan 21 17:09:35 UTC 2015 from open source release
Log: 1 nodes, 16 cpus.
Log: Defaulting to 16 copy threads
Log: Total 16332 MB. Free 13233 MB. Hugepages 0 MB. Targeting 15323 MB (93%)
Log: Prefer plain malloc memory allocation.
Process Error: memalign returned 0
Process Error: failed to allocate memory
Process Error: Sat::Initialize() failed

Status: FAIL - test encountered procedural errors

Process Error: Fatal issue encountered. See above logs for details.

I have no idea whats going wrong. Seems like I'm missing something.
 

·
Registered
Joined
·
4,200 Posts
try

stressapptest -W -s 3600 -M 14000
or
stressapptest -W -s 3600 -M 13000

if you have 16gb

it depends how much memory your os have already allocated

go by this line when it error like that
Log: Total 16332 MB. Free 13233

so you would need to

stressapptest -W -s 3600 -M 13000
 

·
Registered
Joined
·
330 Posts
Quote:
Originally Posted by zGunBLADEz View Post

try

stressapptest -W -s 3600 -M 14000
or
stressapptest -W -s 3600 -M 13000

if you have 16gb

it depends how much memory your os have already allocated

go by this line when it error like that
Log: Total 16332 MB. Free 13233

so you would need to

stressapptest -W -s 3600 -M 13000
Alright thanks. 13000 works. Wish the instruction was more clear for people like me who never used linux distro.
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #632
Quote:
Originally Posted by Worldwin View Post

Alright thanks. 13000 works. Wish the instruction was more clear for people like me who never used linux distro.
By allocating 13000 you're not testing enough of the memory. The instructions are clear enough, it's simply that the RAM GSAT is trying to allocate isn't available.

This is why it's recommended to actually install Mint rather than use BASH.
 

·
Registered
Joined
·
4,200 Posts
Quote:
Originally Posted by Silent Scone View Post

By allocating 13000 you're not testing enough of the memory. The instructions are clear enough, it's simply that the RAM GSAT is trying to allocate isn't available.

This is why it's recommended to actually install Mint rather than use BASH.
you are not saying that plus you also have to understand every system is different.

Its like me saying every test run here so far with page file on is not valid ...
thumb.gif

When stress testing a system first thing you need to do is turn off the page file and do tests in a clean boot. That dont guaranteed you you are going to have a benching/os free resources.
XD
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #634
Quote:
Originally Posted by zGunBLADEz View Post

you are not saying that plus you also have to understand every system is different.

Its like me saying every test run here so far with page file on is not valid ...
thumb.gif


XD
I'm saying it now, I can say it twice if you'd like? Given the difficulties users faced with X370, expecting them to install Mint was a stretch, so I introduced the option of using BASH.

You'd need to elaborate on why you think it's the same as running the test with page file enabled. Depends if you've been allocating memory correctly when testing.
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #635
Your edit is still not relevant. If you're testing memory, you want to be allocating more than 80% as is instructed in the OP. If you don't feel that's right, feel free to start your own memory thread where the actual testing of memory doesn't take place
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #637
Quote:
Originally Posted by zGunBLADEz View Post

you need to allocate as much as possible i do agreed the problem is you dont know if windows is throwing some into the page file <===

If you turn off page file you have more control over this..
It's not difficult, assuming you're not using the machine (which you should not be) whilst the test is running, allocate 90% of what is available, which gives the OS room to breathe. There's no need to disable page file if done correctly.

The problem is allocating 13000 is not a thorough test. As is indicative by the fact GSAT is automatically trying to allocate more. Simple really.
 

·
Registered
Joined
·
4,200 Posts
Quote:
Originally Posted by Silent Scone View Post

It's not difficult, assuming you're not using the machine (which you should not be) whilst the test is running, allocate 90% of what is available, which gives the OS room to breathe. There's no need to disable page file if done correctly.

The problem is allocating 13000 is not a thorough test. As is indicative by the fact GSAT is automatically trying to allocate more. Simple really.
Thats the problem your command is not going to work for everybody as you see right now.

is either you get rid of the page file, do an outside test no os running.. Or you have to deal with the memory allocation
 

·
Overclocking Enthusiast
Joined
·
11,378 Posts
Discussion Starter #639
Quote:
Originally Posted by zGunBLADEz View Post

Thats the problem your command is not going to work for everybody as you see right now.

is either you get rid of the page file, do an outside test no os running.. Or you have to deal with the memory allocation
Or you stop over thinking it, install Mint as suggested and allocate a more suitable amount to the test.

As already said, it's not difficult, but allocating less memory to the test is not the answer. Nothing else need some to be said on the matter. If you want to do things half baked, that's obviously your decision.
 

·
Registered
Joined
·
4,200 Posts
Quote:
Originally Posted by Silent Scone View Post

Or you stop over thinking it, install Mint as suggested and allocate a more suitable amount to the test.

As already said, it's not difficult, but allocating less memory to the test is not the answer. Nothing else need some to be said on the matter. If you want to do things half baked, that's obviously your decision.
but thats not the case tho i purposely turned the page file off to make sure im not testing the page file that defeats the purpose of testing the ram to begin with right?

Both hci and sat tests i have done and share here are page file free/off pc have barely no ram for herself left...

The problems is if you dont tell the gsat what memory to allocate is going to error out you can empty the ram, but you cant take the os ram usage out you can lower it, but theres so much you can do.

The problem is gsat eventually is going to error out everytime if he dont get the automatic 93% of the TOTAL ram amount.,. Like the user above.. He would need to close everything i mean EVERYTHING and try if that doesnt work he would need to allocate manually the ram

Plus if we see

HCI recommended by post as suggested
Quote:
16 instances with 850MB per instance.
that equals to 13,600mb ram

I have instances of hci running 880 each in my settings with no page file

this is gsat on my pc
Quote:
[email protected]:~# stressapptest -W -s 3600
Log: Commandline - stressapptest -W -s 3600
Stats: SAT revision 1.0.6_autoconf, 64 bit binary
Log: buildd @ kapok on Wed Jan 21 17:09:35 UTC 2015 from open source release
Log: 1 nodes, 16 cpus.
Log: Defaulting to 16 copy threads
Log: Total 16336 MB. Free 13202 MB. Hugepages 0 MB. Targeting 15327 MB (93%)
Log: Prefer plain malloc memory allocation.
Process Error: memalign returned 0
Process Error: failed to allocate memory
Process Error: Sat::Initialize() failed

Status: FAIL - test encountered procedural errors

Process Error: Fatal issue encountered. See above logs for details.
[email protected]:~#

Dont matter what hes trying to allocate that 93% ram same as the other user

My windows 10 is bare minimum as it is XD

i can try

stressapptest -W -s 3600 -M 14000

or the same settings as hci

stressapptest -W -s 3600 -M 13600

i try 14000 first theres no way i will have 7% ram usage on a clean boot

closed chrome memory hogger and still ..
Quote:
[email protected]:~# stressapptest -W -s 3600
Log: Commandline - stressapptest -W -s 3600
Stats: SAT revision 1.0.6_autoconf, 64 bit binary
Log: buildd @ kapok on Wed Jan 21 17:09:35 UTC 2015 from open source release
Log: 1 nodes, 16 cpus.
Log: Defaulting to 16 copy threads
Log: Total 16336 MB. Free 14743 MB. Hugepages 0 MB. Targeting 15327 MB (93%)
Log: Prefer plain malloc memory allocation.
Process Error: memalign returned 0
Process Error: failed to allocate memory
Process Error: Sat::Initialize() failed

Status: FAIL - test encountered procedural errors

Process Error: Fatal issue encountered. See above logs for details.

usually -m 14000 do the trick XD
 
621 - 640 of 5072 Posts
Top