DjinnKiller
-
Posts
130 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Posts posted by DjinnKiller
-
-
That game engine is dung heap of epic proportions, and it can't utilise the hardware in any efficient way.
And if it wasn't for all the absolutely brilliant modders on Nexus, I would have ditched this game a very, very long time ago.
Btw, I've drastically cut down on my load times by binding the fps limiter in my ENB to a key. When the load screen starts I push the key to remove the limiter (thanks to a tip from RedRocketTV), the fps goes up to anything from 400 to 900 fps and the load screen takes 3-5 seconds. In heavy areas, maybe 10-15 secs....huge difference.
When the loading has finished I enable the limiter again.
Doesn't solve the fps problem in-game, but at least no more endless waiting...... :)
-
I don't think that this is how it works.
And You would be correct, in the title screen the engine is essentially "freewheel burning". That number, while funny to look at, has little implication on the game.I have 1200+ fps in the title screen, I still get low thirties in some bad areas (Trinity Tower anyone) with full mod loadout. -
Thank you very much for the update.
Sorry to hear it didn't fix your problem, that is a bummer.
And, thanks for the links, they are stored for future reference. :)
-
I've done both and haven't noticed anything either way. But (my usual disclaimer) ymmv. Just make a backup, test and if it goes pear shaped, roll it back.
-
The load time is in some insane way also connected to the FPS. I guess this is what the accelerator does trickery with?
I use an ENB to lock it. Did a bind of the enable/disable (thanks to a tip from RedRocketTV). Before loading I switch off the limiter, the FPS goes up to 8-900, after loading I enable the limiter.
The difference in load times with and without limiting the FPS is pretty staggering.
(At least in my game, ymmv and all the usual stuff.....)
-
Interesting.
Please keep us posted after you apply the fix, if you wouldn't mind!
-
Wise words. Don't overclock if you don't know why you should!
I can't imagine there is anything noticeable to gain with FO4 using XMP.
For me it actually was the other way around, I run some workloads that put heavy load on the CPU and memory (and the Threadripper likes fast memory). The overclock is 100% stable.
But, at the end of the day when I was ready for a beer and mess around in the wasteland, it all went pear shaped.
That's why I asked the OP about XMP.....
-
Fans are a dime a dozen. It is the same manufactures that makes non-branded and branded ones.
If having fans branded with Cooler Master, Corsair etc gives you a warm and fuzzy feeling inside :) by all means, buy those. If cost is an issue, buy non-branded ones. If noise is an issue, use the largest fans possible.
Fans are basically......fans. Don't make this into some kind of art.
Thermal paste is the same, unless you want a metallic one, then there are considerations regarding the metal used in the cooler. If you de-lid the cpu you are in another complexity hell altogether.
-
Cool that you got it sorted. As to why XMP fcuks shite up, who knows.
My newest build is a Threadripper 2950, G.Skill 64GB, RTX 2080 Ti system.
When I flip on XMP profile 2 (3200 MHz) I can load the system very heavily, have screenshots where the mem use is above 62GB and the average load across all cores is above 70%, and everything is rock stable.
...then, reboot the system, start FO4, and it all goes to hell...... switch off XMP and FO4 behaves.
Edit: Btw, SLI is fine. My old system with dual GTX 980's was a blast. But the relatively minor increase in performance does not justify (in my case) the cost of two cards.....
-
It may not make any difference at all, as said, I have no clue why this happens on my PC, and haven't had time to try to find the cause. Built it my self so I know the hardware, but not why it is such biatch regarding FO4 and XMP! :)
-
Try switching off XMP and see if makes any difference. It's a matter of timing (I think), not faulty RAM (Memtest).
I can run my RAM with the second XMP profile (3200 MHz) and all plays along nicely, I can max out my memory and really put load on the CPU, no errors. Then firing up FO4 and things become very unstable.
I don't know why, haven't dug into it, but switching off XMP (down to 2666 MHz if I remember correctly atm), and there is no problem.
-
Bad Memory stick?
...or a XMP profile activated in BIOS?
-
thanks everyone for the comments, i used a can of compressed air a couple of days ago, and it still crashes heaps, but the crashes have been less, so either it's a dust/ventilation problem or something similiar, i have a front fan on the tower, and one on back but the predator sense only see's the front one, how much would it cost to get another fan for it, maybe one on the side of the case, ??
thanks again for the responses
You need to get some facts on what is actually happening at the time of the crash (assuming this is hw/temp/something related and not a sw problem).
Download GPU-Z (link to the portable version), got to the Sensor tab, set it to log to file, and then let it run until the crash occurs.
https://portableapps.com/apps/utilities/gpu-z-portable
Download Generic Log Viewer, and load the log file from GPU-Z, see if the sensor data show anything strange, just before and up until the crash
https://www.hwinfo.com/add-ons/
If you want to go full bore, download HWiNFO (link to the portable version), start it in sensor view only, log to file, and then let it run until the crash occurs.
https://portableapps.com/apps/utilities/hwinfo-portable
Open the log file in Generic Log Viewer (it supports GPU-Z, HWiNFO and AIDA 64) and look for anomalies.
-
As for Windows, I rarely have BSOD's, when I do they always dumps into two categories.
1. Hardware, either me doing something stupid regarding overclock, or some hardware that is faulty.
2. Some driver or software that runs in admin context (RAID-drivers in particular, or buggy security software).
Btw, I run Windows Defender with default settings, never touched it, and it doesn't cause trouble, for me at least, ymmv!
If you have a Corsair with a Link-module (hw thingy), you can use the Corsair Link software to read some metrics from the PSU.
But, what is the BSOD-error, especially since it only happens when playing FO4. I'm starting to get curious! :)
Good luck!
-
it seems to work fine with most other games, but it just simply hates anything from bethesda, it constantly overheats, and throws me out constantly i have tried turning up the fan to full but it just doesnt make any difference, it will get to 60-65c and then boot me out, i know it's not a mod, it's the game, it's too resource heavy...
How is 60 to 65 overheating? Mine gets up to the high 70ies without any problems. And I haven't got the strongest horse in the race with a 650 GTX.
There has to be something else causing you to CTD.
I'm with cossayos on this one....
A stock 1060 should:
- start the fans at around 60C
- start thermal throttling at around 83C
- shut itself down at around 92C to prevent damage (it can usually handle higher temps, but that is the safety limit)
My card (not a 1060, but Nvidia) fluctuates at around 70-75 under heavy load.
-
Weird stuff.
I've been using PRC for a long time, but have tried a lot of different ENB's and never had problems like you describe.
Got some problems with an ENB a while ago, but that was tied to the version, think I had to roll back to 0.307 or something.
After roll back everything was working fine.
Currently using 0.355, and it has been un-problematic. Gonna have a look at the newest version soon.
-
Cool that you got it sorted.
I've never had any problems with ENB, and I use ENB all the time.....strange!
But thanks for thorough explanation.
-
Not an answer to your question I'm afraid (you said nothing about price range, I have a RTX 2080 Ti, it works excellent but costs a small fortune....).
But something seems seriously wrong with the setup.
The 1060 starts thermal throttling at 83C, and the fans starts at around 60C (if I got my specs right).
You sure that it's not something that can be fixed, like airflow and such?
-
First of all, apologies to the OP for hijacking this thread and going off-topic....
@Di3sIrae
The tool you keep posting links to is nothing more than a CLI to some memory management functions in the core OS.
There is no clever programming going on, there is no magic.
For example flushing the Working Set of a process does nothing more than create a lot of page faults as the OS has to swap (parts of) the information back in.
I suggest downloading VMMap from Sysinternals, attach it to a process, clear the working set and watch what happens.
When there is a problem with FO4 (or possibly hardware) you don't start to randomly mess around with the memory management of the OS. You methodically troubleshoot the problem.
As for the ENB-setting, for those who got it working, good for them. I have not, on two very different hardware configurations, and after many hours of testing different values.
Why 4GB (4096) and 13.7GB (14000) should be considered "normal" values I have no clue at all.
-
@RadoGamer
You don't have "too much RAM", there is no such thing as too much RAM.
As I said in my previous post, I run FO4 on machines with 32 and 64 GB of RAM.
Rest assured that this is not the issue.
@Di3sIrae
Stop spamming about "EmptyStandbyList".
In the same paragraph above you mixes up VRAM use and offloading to RAM in the context.
"Normal size like 4096 or 14000", this is WRONG.
4 096 MB of RAM = 4GB, 14 336 MB of RAM= 14GB
I've done A LOT of testing with the setting in enblocal, it has no effect what so ever.
These tests was done with dual GTX 980 on a 32GB system, and RTX 2080 Ti on a 64GB system.
-
This may help if you are low on physical memory, but in general those kind of tools are snake oil.
It's like some years ago when you could buy "memory optimizing" software. All they did was flushing the working set or some other nonsense. There is no magic when handling memory in an operating system.
Microsoft knows what they are doing and the memory management in Windows 10 is very efficient except for some very specialized workloads... Rule of thumb, don't mess with the memory manager, that goes for Windows, Linux, BSD, Unix, z/OS etc.....
Jennifur68 has 16GB of G.Skill TridentZ DDR4 3200mhz RAM, so I would advise to not fcuk around with bullshite tools like this.
Jennifur68 also have a SSD so thrashing is a non-issue....
-
Seems to me that you have covered most of the bases.
My fps fluctuates wildly, from 80 fps (locked with ENB) down to 25-30ish... (RTX 2080 Ti)
But it always climbs back up in less "noisy" areas.
Don't know if this helps you, but a couple of things to try...
1. Lock fps with an ENB
Set iPresentInterval=0 in BOTH Fallout4Prefs.ini and Fallout4.ini MANUALLY.
(I'm paranoid so I also change it in Fallout4_Default.ini.
Lock fps in enblocal.ini
Launch FO4 with a mod manager, do not use the FO4 Launcher, it screws with iPresentInterval=0 in the ini's
2. Run metrics collection before you start FO4
Use HWiNFO, https://www.hwinfo.com
Start it in sensor mode and log values to a file.
You can then read and compare sensor data with Generic Log Viewer, https://www.hwinfo.com/add-ons/
Wait until the drop occurs, play for another 5-10min, then exit the game.
Check things like proc load, GPU load, VRAM load, temp (is it thermal throttling that screws with your game?) etc. in the file created by HWiNFO in the time interval just before the drop and forward.
As I said, haven't a clue if this brings you anywhere, but I would start there......
Happy hunting...
EDIT:
Since you said "about 30-40min", thermals may be a cause.
Windowed mode always drops my fps significantly, borderless or not, that is my experience on both GTX980 and RTX 2080 Ti.
The windowed borderless "fix" always makes things worse on my rigs, regardless of other settings.
-
Fallout 4 works fine on Windows 10 Pro.
I've run it on two diferent PC-builds, one with an old Intel i7-5960X Extreme, Corsair Dominator DDR4 2800MHz (2x16GB), and dual MSI GTX 980, and another with AMD Threadripper 2950X, G.Skill Flare X DDR4 3200MHz (8x8GB) and Gigabyte RTX 2080 Ti.
Had my fair share of crashes due to faulty hardware and software related shite through the years, but when those are sorted, FO4 and other games works fine on Win10 Pro.
Your description of the problem is kind of non-specific, not much hard facts to go on here....
1. What is the BSOD error code & error text
2. Did you log sensor metrics over time and look at them at the time of the crash?
I personally prefer HWiNFO, https://www.hwinfo.com
Starting it in sensor mode and log values to a file.
You can then read and compare sensor data with Generic Log Viewer, https://www.hwinfo.com/add-ons/ after a crash
3. Did you have any abnormal messages in the Windows Event Log (Application, Security and System) at the time of crashing?
4. Do you have any third-party anti-malware sw, or sw that periodically executes and that coincides with the crashes?
5. Do you have a PSU that can log data to file (input/output voltage, rail voltage, efficiency, temperature etc), for example like the Corsair AX1200i?
6. Do you overclock the system or any parts of it?
7. Did you log data from Windows Performance Monitor over time, especially the Memory, Network Interface, PhysialDisk and Processor counters?
8. ....and if you want to take it a step further you can enable process tracking in the Windows Event Log (Security),
or take a look at NirSoft's ExecutedProgramsList or similar tools to get further insight into what is happening at the time of the crash.
Point 2-8 is not primarily to pinpoint the cause, but to rule things out.
Happy hunting!!
EDIT: Process Monitor from Sysinternals may also be helpful.
-
Happy to help, cool that it worked out!
Fallout 4 lockup (completely freezes)
in Discussion
Posted
They didn't.
The case (back around 2016) was much more convoluted with a mix of Win8, Win10, DX9-games, 32-bit vs 64-bit, potential driver issues (both Nvidia and AMD), and so on and so forth.
Ie. the usual confusion and FUD on this beautiful thing called the Internet.... :)