Discussion in 'Videocards - AMD Radeon Drivers Section' started by adrock311, Apr 3, 2011.
my MSR values are the same but I saw an increase. Depends on cpu, I guess.
You should change your first post in this thread to reflect this, as to not confuse people
EDIT: NM you didn't start this topic
No difference here after trying the TLB Disable patch.
One question, does this make any permanent changes to the system?
From what I gather, once you reboot things are back to normal?
Yeah once you reboot things are back to normal
The settings are only memory address changes, they don't change the systems registry or anything
Ok thanks for that.
Yea I would have if I did, thats why I posted this last post to clarify my mistake
Testing tlb on phenom 2 955 be
Ok guys they say not affected phenom 2 my ass look at this result
running:win 7 32bit sp1 unlocked 4gb memory to use
amd phenom 2 x4 955 be stepping 3 agena clocked at 3,616ghz
motherbard formula 2 am2+
tested on unigine haven default settings.1680x1050
min fps 8,1
max fps 78,9
min fps 16,7
max fps 78,6
look at minimum fps its doubled
i hope this **** will be fixed in next version of windows
Sounds just like a fluke in benchmarking to me
I've had cases with Dirt 2 benchmarking where average scores were normal, but minimum were only 7-10FPS below it, and I run the benchmark 5 more times, and the minimum is then on average 20fps below it
Never just run ONE benchmark and then conclude the results
Also, your total score only went up 3 points, which sounds within the norm percentage fluke of benchmarking, to me that sounds like TLB patch was already disabled with your Phenom II
Get serios man.This is 3rd run on all.system default/tlb disable.All peapole in the world know that 1st run benchmarking scores is not accurate.I talking about minimum fps where is quite difference.
And I am telling you that you have a fluke in results
Look at your max FPS with TLB disable, its lower then before you even applied it!
You would be seeing a increase of 30% across ALL scores of TLB disable was working, not just a fluke on the minimum FPS
And also, you are using the wrong software for testing, you should be running something like superPI before you apply the patch, and then run it after the patch, and find out your CPU results
You are testing your videocard, not your CPU, and like I said I have had fluke results with videocard testing before where 5+ runs are where min FPS is 20 below average, then that one run where min FPS is 6-7FPS below average
Download Sisoft sandra and do a CPU benchmark, run it once after boot, then once after applying the disable patch and see if you get wildly different results, cause right now you are testing your GPU, not your CPU
EDIT: Looking at your post, you are running uiengine, which is a videocard benchmark, not a CPU benchmark
I had the same results in unigine benchmark, min fps more than doubled with tlb disable, and also less framerate drops in many games, i doubt its a fluke
Again you need to run a benchmark like SuperPI and check your results
Reboot your system and run superPI and record the results, then apply the disable patch and record your results, if they are WILDLY different then yes you have the TLB enable bug still being applied with your windows installation
You notice with when I made that thread 2 years ago with this very issue I only used CPU benchmarks and not GPU benchmarks?
Now you are talking nonse ofcourse its for gpu testing but also it uses cpu power.run this bench on same system with older cpu and you will see result
Captain Janeway OUT!
TLB erratum error is a CPU thing, not a GPU thing
God, screw it, I'm just going to call you a -MORON- until you freaking do a CPU benchmark properly
Download SuperPI and unzip it, then reboot your system, run superpi and do the 2m test, find out your score, then do TLB disable patch and then run superpi and 2m test again
If SuperPI completes the test FASTER after TLB disable patch is applied, congrats, your CPU is working faster, if it -DOESN'T- go any faster, then you aren't affected with it and windows
That goes for all of you, running GPU benchmarks to find out CPU scores is freaking retarded, CPU's are nowhere near taxed enough running something like uniengine etc
Hell Vantage -HAS- a CPU test in it, go run that and see if your scores are a lot higher after using the TLB disable patch. First reboot into windows, run vantage, then apply the TLB disable patch and run Vantage again, if your scores go way up, then you know its working
Cheers.. I wondered wtf that setting was in my bios ...
CPU: Phenom II X4 965 BE @2200Ghz_0.8500v
RAM: 1GBx4 800 Mhz DDR2, 4-4-4-12
GPU: AMD 6950
OS: Windows 7 32-bit
Running the TLB patch disabler, actually decreased performance for my setup.
1. Memory read speed decreased by nearly 500 MB/s
2. Memory Latency increased by 8.0 ns
3. L3 Cache Read speed decreased by 1000 MB/s
4. L3 Latency increased by 2.0 ns
Ahem..gonna restart and check up speeds again..
Hmmm my BIOS has a read-only info about Microcode something, but there is no setting associated with it. Is there any hidden super duper extra key to show settings in a BIOS like mine?
It could just be random benchmarking variances too
CAUTION: not all benchmarks indicate TLB enable/disable performance change equally:
check out the DIFFERENCE column to see which benchmarks are effective for testing TLB bug.
another site about various benches to test TLB:
again, check out the difference column there.
Yepp speeds in my benchmark actually decreased since the TLB Patch Disabler..
The following are some benchmarks!!!.
800Mhz@0.7000v Left side image(Default) | Right side image(With TLB Patch Disabler)
2200Mhz@0.8500v Left side image(Default) | Right side image(With TLB Patch Disabler)
2700Mhz@0.9875v Left side image(Default) | Right side image(With TLB Patch Disabler)
3400Mhz@1.2000v Left side image(Default) | Right side image(With TLB Patch Disabler)
hmm..any ideas why!..