I forgot to mention that I can use Radeon Pro as a workaround to set a Frame Rate Limit instead of attempting to use Crimson Frame Rate Target Control when in Crossfire. Radeon Pro Frame Rate Limit seems to work fine with no issues.
I have just completed some extensive testing on "The Division".
I checked my system is clean and is running the latest updates.
I turned off all overclocking on CPU and GPU.
I ran the following tests with all Graphics settings on "The Division" set to maximum possible settings at 1080p.
I did not set NVIDIA Algorithms for Ambient Occlusion or Shadow Quality in case these algorithms cause issues with AMD cards.
In game FPS limit set to 60FPS initially but also tested turning it off.
I have tried With Default Crossfire and "The Division" crossfire profile.
1. Frame Pacing = on and Frame Rate Target Control = on. Game runs for anywhere between 1 to 10 minutes then my entire PC freezes and I have to reboot.
2. Frame pacing = on and Frame Rate Target Control = off. Game runs for anywhere between 1 to 10 minutes then my entire PC freezes and I have to reboot.
3. Frame pacing = off and Frame Rate Target Control = on. Game runs for anywhere between 1 to 10 minutes then my entire PC freezes and I have to reboot.
4. Frame pacing = olf and Frame Rate Target Control = off. Game runs really well. I get 60FPS mostly. It works great.
CPU use is high with Crossfire enabled. It sits in the range of 90-100% when I run around in the game.
Any Ideas / potential solution would be great.
No one has this issue?
I filed another report using AMD Online Reporting form regarding this.
Update. I just purchased a Second Hand R9 Nano and I am now testing it in the same System.
I removed the Sapphire HD7970 OC 6GB Vapor-X and Sapphie R9 280X OC 3GB Tri-X Vapor-X.
Borderlands 2, runs fine in Crossfire the above Crossfire Setup (provided I keep Frame Pacing and Frame Rate Target Contol switched off) with these Crossfire Cards.
Using a single R9 Nano in the above system and running Borderlands 2 with exactly the same graphics settings again runs fine until I turn on Frame Rate Target Control. If I turn on Frame Rate Target Control with the R9 Nano, I see the same symptoms as previously with my Crossfire Setup.
My PC Freezes after 5-10 minutes of gameplay.
I am using freshly installed Radeon Software Version 16.11.4 on a Clean Windows 8.1 System (Checks were done and are O.K.) and Borderlands 2 files have been verified.
Therefore it looks like ther is a PC Freezing Issue with Frame Rate Target Control on:
1. Crossfired HD7970/R9280x - Turning Frame Pacing on, Frame Rate Target Control on, or both on cause my PC to Freeze on multiple titles.
2. Single R9 Nano. - Turning Frame Rate Target Control on causes my PC to Freeze on Borderlands2, more testing to follow.
I will go on to test what happens with the R9 Nano on other titles
I do not have time to do that this weekend though.
I do not have a second R9 Nano to test Crossfire Performance yet. The supplier of the card let me down. I am looking for another.
I have been reporting this Frame Pacing / Frame Rate Target Control Issue for a year now.
I understand that AMD may not have had time to look at my Sapphire HD7970 OC 6GB Vapor-X and Sapphie R9 280X OC 3GB Tri-X Vapor-X Crossfire issues.
I understand newer cards would be a priority.
But now I am seing Frame Rate Target Control PC Freezing with a Single R9 Nano.
I have tested every other feature of this R9 Nano apart from a Full 4GB memory check.
Everything else is working like a dream on the Nano.
Freesync is amazing. Tom Clancy "The Division" runs fully maxed out at > 60FPS at 1080p.
I think the Nano is a fantastic card, especially at the price point I bought it for.
Has anyone out there seen this PC Feezing issue with the R9 Nano and Frame Rate Target Control?
Perhaps it is just me ...
Are you trying to play with both Freesync and Frame Rate Target Control enabled?
Try to close (unload from system) all programs like "Radeon Pro" (and reset all tweaks in those programs to defaults before closing)
It is highly recommended to uninstall all AMD software and clean your system in safe mode before swapping your video cards.
Sapphire HD7970 OC 6GB Vapor-X + R9 280X OC 3GB Tri-X- not the best choice for perfect behavior in crossfire.Especially with Freesync or Frame Rate Target Control enabled.
Thanks for yor response.
1. RE: Are you trying to play with both Freesync and Frame Rate Target Control enabled?
Freesync is disabled in AMD Radeon Settings in all tests regarding Frame Pacing on my system.
I wanted to compare the results as best as possible under the same test conditions.
HDMI output was used from the Nano.
Freesync is not supported with the Sapphire HD7970 OC 6GB Vapor-X and Sapphie R9 280X OC 3GB Tri-X Vapor-X.
I use Sapphire HD7970 OC 6GB Vapor-X as my Primary card in the Crossfire Tests.
I tested the DVI& miniDisplay & HDMI outputs from the Sapphire HD7970 OC 6GB Vapor-X in ther Crossfire System.
2. RE: Try to close (unload from system) all programs like "Radeon Pro" (and reset all tweaks in those programs to defaults before closing.
I was not using RadeonPro during these tests.
RadeonPro is installed on my system.
I run the following tests:
(2A). Repeat the test by opening RadeonPro. Resetting it to default. Closing. Repeat test.
(2B). Repeat the test by uninstalling RadeonPro and anything else like it (MSI Afterburner for example). Running clean uninstall. Reinstall. Repeat test.
3. I will try running DDU in Safe mode and repeat the test.
Again, thank you for the suggestion.
4. RE: Sapphire HD7970 OC 6GB Vapor-X + R9 280X OC 3GB Tri-X- not the best choice for perfect behavior in crossfire. Especially with Freesync or Frame Rate Target Control enabled.
I do understand that the cards are not on the official AMD Crossfire Compatibility chart.
I wanted two perfectly matched Sapphire HD7970 OC 6GB cards but could not get hold of a pair of these at the time.
I worked out that since the R9280x was still using Tahiti chip it should be OK to Crossfire the Cards. This was since confirmed from AMD Red Team folk.
The difference in Memory should not really be an issue because they both only use the first 3GB in Crossfire due to the 6GB versus 3GB difference in both cards.
Not sure if the Spare 3GB in the 6GB HD7970 can be used for anything when in Crossfire with a 3GB card. I believe it is wasted if I am running on 1 monitor.
DX12 multiGPU behaviour may be different in the future.
Freesync is not enabled or supported for these cards as far as I understand it from AMD info on FreeSync.
I do have Mini DisplayPort outputs on the Sapphire HD7970 OC 6GB.
I *think* it came with a Mini DisplayPort to DisplayPort Adapter,
I will need to check the DisplayPort output specification for the Sapphire HD7970 OC 6GB to see if it might actually work with FreeSync.
The fact that the system freezes after 5-10 minute of gameplay with Frame Rate Target Control (and/or Frame Pacing for that matter) on these Crossfired cards is
annoying but I do understand they are perhaps an unusual combination. I also understand that AMD would be concentraring on newer cards like the Fury and Polaris and hopefully Vega Cards.
If I perform the above tests that you suggested and I still see a problem with Frame Pacing on the single Nano I will report the issue to AMD.
I just managed to source another (Brand New) XFX R9 Nano 4GB for £300 from Amazon UK.
I needed some more AMD Cards for OpenCL work I am doing.
I was going to hold out for Vega but I needed more horsepower today.
I was so impressed with the performance of my "second hand" PowerColor Nano I decided to purchase another.
At least my money did not go to Nvidia eh? ;-).
I will now be able to Crossfire two Nanos to see how they perform.
I should also be able to look at FramePacing and FrameRate Target Control on a pair of Crossfired R9 Nanos.
In addition I will be able to look at Crossfiring over thunderbolt ... just for a laugh you understand.
My new XFX R9 Nano should be here by the weekend.
I will post the results as soon as I can.
It was good to get a response from someone.
It can get lonely out here talking to myself sometimes ;-).
OK I just received my 2nd (Brand New) XFX Nano.
I can now test R9 "Nano Nano" Crossfire Performance.
Someone wanted me to look into that on another thread anyhow.
New drivers have just been released:
Radeon Software Version 16.11.5
So I should do a fresh driver install anyhow.
I did the following:
Uninstalling RadeonPro and anything else like it (MSI Afterburner for example). - None of these programs actually run at start up but I removed them anyhow.
Running clean uninstall as advised.
Power Down PC.
Fit the new XFX R9 Nano to my motherboard.
Install Radeon Software Version 16.11.5.
Restart the PC.
Check both R9 Nanos are seen and Crossfire is on.
I will start testing Frame Pacing, Frame Rate Target Control, Crossfire Scaling next.
Freesync is off for all tests and I will use HDMI out.
Here are initial performance results for "The Division" testing on the Crossfired Nanos below.
Next step is turn all overclocking off. Leave Frame Pacing and Frame Rate Target Control off.
Pick a game level and play it for at least 30 minutes.
Then exit the game.
Then set Frame Pacing on.
Then start the game.
Then see how long it takes before my PC Freezes ... if it does.
Repeat for Frame Rate Target Control and then for both.
Initial results for The Division with a Pair of Fury Nano's and Radeon Software Version 16.11.5
2 of R9 Nano in Crossfire.
Default Settings in Crimson but with Frame Pacing and Target Frame Rate Control off.
HDMI out used.
AMD Freesync off.
Test = The Division Benchmark on Ultra with Vsync off.
Crossfire Profile used = Default.
No Flicker seen during the benchmark.
(1). No overclock (CPU and GPU at Stock) results.
Crossfire off. 49.9
Crossfire on. 61.1
Crossfire Scaling 1.22
(2). CPU overclock of 16% GPU's at Stock Clock Speed but Both GPU Fan and Power Targets set to maximum.
Crossfire off. 53.1
Crossfire on. 69.2
Crossfire Scaling 1.30
I am not overclocking either of the Nanos at the moment because I have to verify they are o.k. following purchase.
I have a "Window" to return them if they fail, on condition I do not drive the GPU clocks beyond stock setting.
The scaling for the Division I achieved with Crossfire last time I checked the Division
was in the range of (1.35 - 1.40) with my Sapphire HD7970 OC 6GB / R9 280x OC 3GB cards.
But for those cards I was overclocking the GPU Clocks and Memory to the very edge of stable operation.
The scaling factor went up the more I overclocked up to my final overclock levels for those cards.
I would obviously like to see better Crossfire Scaling than this on "The Division" of course.
However compared to how Crossfire was behaving on this game at launch, this is a massive improvement.
One point to note. The Total CPU Utilization Use on my System (i74770K) is getting rather high (90-95%) during the benchmark.
I think this game is very heavy on CPU resources. The Division reported Benchmark CPU% does not reflect the total CPU Utilization that I see.
I have managed to see 100% CPU Utilization in game with my Sapphire HD7970 OC 6GB / R9 280x OC 3GB cards in the past.
I will now go back and run some in game testing and report back if I see any flicker.
Division is not the best game to test crossfire. People notice that is no scaling at all in some places of this game. Try Battlefield 4 (or even better BF3) and other games with good scaling.
Yes I understand "The Division" does not scale well with Crossfire.
I will look others that scale much better as well.
I can't say that I am. The only issue I have is that if I run the benchmark then play the game, it falls to 20FPS and stutters like crazy.
What's your pagefile and how much Ram do you have?
I was hoping to wrap this up based on testing of AMD Crimson Relive drivers and submit my "last ever" Frame Pacing / Frame Rate Target Control Report with a pair of matched AMD cards.
I can report that I see the same PC Freezing issue with 16.11.5 driver during initial testing of a few titles, using the R9 Nanos.
I stopped because of AMD Crimson Relive Release testing.
I have just done lots of testing of AMD Relive version and I have no more time left for now.
However given the current reported and seen issues with Crossfire on the AMD Crimson Relive version anyhow I cannot complete
Frame Pacing and Frame Rate Target Control with newly released drivers.
So I wait to see the Crossfire Fixes for AMD Crimson ReLive.
Perhaps someone in AMD might take the time to test Crysis 3 / Borderlands2 / Tom Clancy "The Division"/ Warframe, etc for around 30 minutes of gaming with Frame Pacing set to 61 and Frame Rate Target Control turned on with a pair of R9 Nanos/ FuryX's and see if anything bad happens?
Wrapping this thread up for now until next driver release.
2 of 2 people found this helpful
Since you're seeing the same issue with 2 different cards, that would rule out the GPU's.
What's the motherboard model you are using ? What is the BIOS version ?
Thanks for the response.
I am using ASUS Z87 Deluxe/Dual, which was a pretty expensive high end board, with the very latest BIOS available on the ASUS Website.
The BIOS Version reported by my Motherboard is: Version 2103 and it was downloaded from here: Z87-DELUXE/DUAL | Motherboards | ASUS Global
I can reflash the BIOS and test again if you you think that is really necessary. It was reflashed earlier in the year. I can look into exaclty when. I tried the Reflash just to test this issue. It made no difference at the time I tested it. Also I have swapped my Corsair AX1200i PSU for a new one, and rewired my PC Cabling. Also now I have the R9 Nanos, and I still see the PC Freezing issue after 5-10 minutes issue with the 16.11.5 driver.
I am not attacking AMD or support in what follows.
I do realize I was using "Crazy" HD7970/R9280x in the past, and I do realise there might be something in my setup causing this.
I am just giving you some feedback below.
I reported the above information in the Issue Reporting Forms many times.
I reported all of the details requested I think.
When I file an AMD Report Form, there is no "Ticket" opened so I cannot see if anything has been done on the reports that I submit.
I do get email saying thanks for submitting. I think I have kept most of them. I cannot see the details that I submit have actually been submitted o.k. though.
This is all I see
"Thank you for taking the time to provide us your feedback on the AMD Catalyst Drivers.
Your feedback is extremely important to us and we review all submissions received.
This mailbox is not monitored. Should you have an issue that requires a technical response, please contact AMD Customer Care
AMD Global Customer Care "
I see the comment that people are asking things on the Forum and not Submitting AMD Report Issue forms.
Well, there seems to be a lack of feedback on what you submit, and no way to track what is happening to the Report Issue you take the time to submit.
When I saw that AMD have added the ability for Users of the Crimson Relive Driver to give feedback to AMD my first thought was this.
Why did the effort not go into the Issue Reporting Form, and improving/automating Quality Control Tracking?
Why can't I simply click "Report Issue", and AMD fill in all my System Details, GPU info, CPU Info. I currently have to do all of that manually.
Crimson clearly knows the majority of this information already.
All I would need to do is fill in some text describing the issue and which games are affected, and maybe add my email, but that could be stored.
I should be able to Click Submit, and get a Tracking Ticket Number where I can log in and see if anything has been done, and answer any questions.
Thanks for looking at it now. I do appreciate it.
I did push on early this morning before I started other work and looked again at how Frame Pacing and Frame Rate Target Control are running with the AMD Crimson Relive Driver on a few titles in Windows 8.1 64 bit. I understand it appears there might be some issue with Crossfire on this Driver. I will report what I saw next.
Here is what I found early yesterday morning on a few games I was previously looking at Crossfire for others on other threads on this AMD forum, whilst testing my 2 R9 Nano's.
Again my System is:
Asus Z87 Deluxe/Dual. GPU = Two R9 Nano. CPU = i74770K. OS = Windows 8.1 64bit. I have the top two x16 PCIe slots on the motherboard occupied with the R9 Nanos, the 3rd x16 slot is empty. The Motherboard BIOS setup is such that both the occupied x16 slots are set to Gen 3, which is how I usually run this PC. The Motherboard manual, and the BIOS reports, that this means actually run the R9 Nanos with two PCIe Gen 3.0 x8 interfaces. If I were to occupy the 3rd x16 PCIe with, for example a Samsumg 1 TB NVMe Gen 3.0x4 SSD, my Slots would become PCIe Gen 3.0 x8, PCIe Gen 3.0 x4, PCIe Gen 3.0 x4 which does affect Crossfire Performance in game benchmarking.
I started with a PC System Check which comes out clean. Fresh Driver uninstall, clean, reinstall. I run all tests with ALL GPU overclocking off, and All CPU overclocking off, since overclocking might cause stability issues. All Gaming Graphics Settings were set to Ultra/Highest Possible setting unless otherwise stated.
I tested the following titles to see how Frame Pacing and Frame Rate Target Control were working in Crossfire.
1. BattleField 1. (Origin Access "Vault" Trial Version). aka BF1.
2. BattleField 4. aka BF4
3. Crysis 3.
4. Borderlands 2.
5. The Witcher 3
7. Skyrim SE. (Special Edition).
Results table and comments next.
You don't weant me to post to the forum I guess.
You don't want me to post to the forum I guess.
I will send the Data direcly to ray_m.
What results did you get with those games?
Were you able to determine if the issues were down to the motherboard?
RE: what results did you get with those games:
Sorry I have not been replying to this forum for a while. I have been extremely busy with work.
Crossfire was not working very well with initial Crimson ReLive Version so I stayed on Crimson 16.11.5 driver until recently.
Latest Results. with Crimson 16.11.5.
Although I have tested this issue to death over the past year ... I tried again.
I reflashed and reset my BIOS.
I Updated Windows to the latest Updates.
I ran standard system check to ensure there is nothing wrong with the Windows 8.1 OS.
I ran the CPU and GPU's (i7-4770K & a pair of R9 Nanos) at stock clocks. No overclock.
I uninstalled Crimson 16.11.5, ran amdcleanuputility, reinstalled Crimson 16.11.5.
In the Crimson 16.11.5 driver all of the following titles froze after around 10 minutes of game play if I turned on Frame Pacing and or Frame Rate Target Control.
1. BattleField 1. (Origin Access "Vault" Trial Version). aka BF1. - Crossfire did not enable with the Origin trial version. I bought it to for testing.
Immediately after purchase, Origin runs an update and you can then immediately get Crossfire to enable. FRTC on.
2. BattleField 4. aka BF4.
3. Crysis 3.
4. Borderlands 2.
5. The Witcher 3
7. Skyrim SE. (Special Edition).
All of the above titles frooze after 10-15 minutes with Frame Pacing and Frame Rate Target Control on.
All of the above titles played fine until I ran out of time to test (~1 hour) with Frame Pacing and Frame Rate Target Control switched off.
Note I had to uninstall clean, and reinstall the driver and rerun all the disk checks and 3dMark Firestrike each time to ensure Crossfire was working and test the next game...
Now for the good news.
Latest Results. with Radeon Software Version 17.1.2.
1. BattleField 1 (purchased version).
2. BattleField 4. aka BF4.
3. Crysis 3.
4. Borderlands 2.
5. The Witcher 3.
7. Skyrim SE. (Special Edition).
And also ...
Frame Pacing and Frame Rate Target control appears to be fixed in
It is working and stable when I run with default (no overclock).
It is also working fine if I run with my tested stable CPU+16%, GPUS=5% overclocks.
I believe the above testing methodology has been pretty stringent.
I am convinced it is the Radeon Software Version 17.1.2. that makes the difference.
I am now able to run Battlefield 4 at 4K & 60FPS with all Graphics settings maxed out with these R9 Nanos and this Driver.
Similary with Titanfall.
ReLive Recording is also working great for me.
I have one thing left to check but I am pretty pushed for time.
I kept the Sapphire HD7970 OC Vapor-X 6GB and R9 280X OC Tri-X Vapor-X 3GB cards for use in my other PC.
It is an Asus Z97 board with an i7-4970K CPU. I will see if Radeon Software Version 17.1.2 fixes frame pacing on those cards on that machine.
In any case, for the R9 Nanos in Crossfire with the i7-4770K, this issue is fixed with the Radeon Software Version 17.1.2. driver.
This is honestly the first time I have seen Frame Pacing and Frame Rate Target Control working stable for me in Crossfire on my AMD Cards.
I don't know if someone in AMD Drivers Team actually looked into this for me specifically or if it is a pure fluke but ... in any case thanks!