• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

R9 290(x) crossfired thru PLEX CHIP

Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#1
Is it a good idea to run a crossfire 290 (x) thru PLX chip?

Would there be a performance impact?

I have been searching all over net but all I can find are crossfired 290(x) with X79 mobos........

This is a question from a friend and I am just helping him to find an answer....


his rig z87 sniper 5 with 4770K
 
Joined
Oct 29, 2012
Messages
1,889 (0.99/day)
Likes
742
Location
CZ
System Name TITAN Slayer / CPUCannon / MassFX
Processor i7 5960X @ 4.6Ghz / i7 3960x @5.0Ghz / FX6350 @ 4.?Ghz
Motherboard Rampage V Extreme / Rampage IV Extreme / MSI 970 Gaming
Cooling Phanteks PHTC14PE 2.5K 145mm TRs / Custom waterloop / Phanteks PHTC14PE + 3K 140mm Noctuas
Memory Crucial 2666 11-13-13-25 1.45V / G.skill RipjawsX 2400 10-12-12-34 1.7V / Crucial 2133 9-9-9-27 1.7V
Video Card(s) 3 Fury X in CF / R9 Fury 3840 cores 1145/570 1.3V / Nothing ATM
Storage 500GB Crucial SSD and 3TB WD Black / WD 1TB Black(OS) + WD 3TB Green / WD 1TB Blue
Display(s) LG 29UM67 80Hz/Asus mx299q 2560x1080 @ 84Hz / Asus VX239 1920x1080 @60hz
Case Dismatech easy v3.0 / Xigmatek Alfar (Open side panel)
Audio Device(s) M-audio M-track / realtek ALC 1150
Power Supply EVGA G2 1600W / CoolerMaster V1000 / Seasonic 620 M12-II
Mouse Mouse in review process/Razer Naga Epic 2011/Razer Naga 2014
Keyboard Keyboard in review process / Razer Blackwidow Ultimate 2014/Razer Blackwidow Ultimate 2011
Software Windows 7 Ultimate / Windows 7 ultimate / Windows 7 ultimate
Benchmark Scores cinebench 15.41 3960x @ 5.3ghz Wprime32m 3.352 3960x @ 5.25ghz Super PI 32m: 6m 42s 472ms @5.25ghz
#2
It might screw with frame pacing but other than that it should be perfectly ok.
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#3
It might screw with frame pacing but other than that it should be perfectly ok.

Thanks could you elaborate more? I know for the fact that frame pacing of this cards are thru hardware...... I also thought about that but my own guess, only guess is that theoritically, it should be better because it will be filtered the video rendering thru PLEX chip.... its like a 2 rivers going into single river going to open SEA, (CPU)..... but that is just what I thought. I would want to see thru bench of course or personal experience......

He is also searching and searching though, I have more time searching thru internet.....


edit: He has same monitor resolution as mine 2560 x 1600..... we were actually thinking of buying one more 2560 x 1600 monitor and test if 290 crossfired can be able to handle games in

4800 x 2560 (portrait) or 7680 x 1600 resolution...... but both of us dont have 2011 CPU and mobo..... ^_^
 
Joined
Oct 29, 2012
Messages
1,889 (0.99/day)
Likes
742
Location
CZ
System Name TITAN Slayer / CPUCannon / MassFX
Processor i7 5960X @ 4.6Ghz / i7 3960x @5.0Ghz / FX6350 @ 4.?Ghz
Motherboard Rampage V Extreme / Rampage IV Extreme / MSI 970 Gaming
Cooling Phanteks PHTC14PE 2.5K 145mm TRs / Custom waterloop / Phanteks PHTC14PE + 3K 140mm Noctuas
Memory Crucial 2666 11-13-13-25 1.45V / G.skill RipjawsX 2400 10-12-12-34 1.7V / Crucial 2133 9-9-9-27 1.7V
Video Card(s) 3 Fury X in CF / R9 Fury 3840 cores 1145/570 1.3V / Nothing ATM
Storage 500GB Crucial SSD and 3TB WD Black / WD 1TB Black(OS) + WD 3TB Green / WD 1TB Blue
Display(s) LG 29UM67 80Hz/Asus mx299q 2560x1080 @ 84Hz / Asus VX239 1920x1080 @60hz
Case Dismatech easy v3.0 / Xigmatek Alfar (Open side panel)
Audio Device(s) M-audio M-track / realtek ALC 1150
Power Supply EVGA G2 1600W / CoolerMaster V1000 / Seasonic 620 M12-II
Mouse Mouse in review process/Razer Naga Epic 2011/Razer Naga 2014
Keyboard Keyboard in review process / Razer Blackwidow Ultimate 2014/Razer Blackwidow Ultimate 2011
Software Windows 7 Ultimate / Windows 7 ultimate / Windows 7 ultimate
Benchmark Scores cinebench 15.41 3960x @ 5.3ghz Wprime32m 3.352 3960x @ 5.25ghz Super PI 32m: 6m 42s 472ms @5.25ghz
#4
The thing with PLX chips is that they don't make more lanes they do traffic management to make it look like there are more lanes. So traffic on the PCI-e goes form this:

GPU1 >> PCI-e 3.0 x8 >> CPU >> PCI-e 3.0 x8 >> GPU 2

to this:

GPU1 >> PCI-e 3.0 x16 >> PLX >> PCI-e 3.0 x16 >> CPU>> PLX >> PCI-e 3.0 x16 >> GPU 2

and according to the PLX website every pass through the PLX chip takes up another 110 nano seconds. In the diagram there are 2 passes through the PLX, off setting frame times by an extra 220ns which is such a big deal so you should be fine.
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#5
ok ok... i know about PLX chips...... i do know also that they dont make more lanes......


what I am after now is that the R9 290 (non X) and R9 290X use PCI-E lanes to communicate each other.....

Theoretically, There is a big performance impact since now if there is PLX chip involve. Imagine like this........

2 person (2 R9 290) calling a company (CPU). but company uses a HUB (PLEX Chip). Unlike with no HUB, it can directly contact company....

Before, GPU's are communicating thru Xfire Bridges....

as to performance impact? Nobody knows unless somebody will really do extensive test.


as for PLEX chip, 220 nano seconds doesnt impact really on game performance.

Anandtech did an extensive test.

http://www.anandtech.com/show/7189/choosing-a-gaming-cpu-september-2013/6

look at Metro 2033. I7-4770K, 3 7970's using PLX has atleast 2 more FPS compared to 3930K, 3 7970's........

I know i know the games are old but that is just an example........


7970's uses bridge, R9 290's using PCI lanes........

That is what I am after for now.....
 
Joined
Oct 29, 2012
Messages
1,889 (0.99/day)
Likes
742
Location
CZ
System Name TITAN Slayer / CPUCannon / MassFX
Processor i7 5960X @ 4.6Ghz / i7 3960x @5.0Ghz / FX6350 @ 4.?Ghz
Motherboard Rampage V Extreme / Rampage IV Extreme / MSI 970 Gaming
Cooling Phanteks PHTC14PE 2.5K 145mm TRs / Custom waterloop / Phanteks PHTC14PE + 3K 140mm Noctuas
Memory Crucial 2666 11-13-13-25 1.45V / G.skill RipjawsX 2400 10-12-12-34 1.7V / Crucial 2133 9-9-9-27 1.7V
Video Card(s) 3 Fury X in CF / R9 Fury 3840 cores 1145/570 1.3V / Nothing ATM
Storage 500GB Crucial SSD and 3TB WD Black / WD 1TB Black(OS) + WD 3TB Green / WD 1TB Blue
Display(s) LG 29UM67 80Hz/Asus mx299q 2560x1080 @ 84Hz / Asus VX239 1920x1080 @60hz
Case Dismatech easy v3.0 / Xigmatek Alfar (Open side panel)
Audio Device(s) M-audio M-track / realtek ALC 1150
Power Supply EVGA G2 1600W / CoolerMaster V1000 / Seasonic 620 M12-II
Mouse Mouse in review process/Razer Naga Epic 2011/Razer Naga 2014
Keyboard Keyboard in review process / Razer Blackwidow Ultimate 2014/Razer Blackwidow Ultimate 2011
Software Windows 7 Ultimate / Windows 7 ultimate / Windows 7 ultimate
Benchmark Scores cinebench 15.41 3960x @ 5.3ghz Wprime32m 3.352 3960x @ 5.25ghz Super PI 32m: 6m 42s 472ms @5.25ghz
#6
The 2FPS difference between the 3770k and 3930k is the fact that a 3770k is faster in single threaded tasks and has nothing to do with the PLX chip. Basically the PLX chip will be slower than native PCI-e by XXX nano seconds. In Xfire the PCI-e bus is used for shipping rendered images from the second card to the first and for the first to tell the second when to start working. The moving of a raw 1080p image over PCI-e 3.0 x16 takes 377ns from the second card to the cpu and another 377ns from the CPU to the first GPU. With the PLX you just add the delay that the PLX takes to re route and add it to both the GPU2 to CPU and CPU to GPU1 transfer and you get the effect of the PLX chip on the frame rate. The PLX chip on your mobo takes 126ns to reroute so it adds 252ns to the time between new frames which is so little that you won't ever notice.
So Xfire r9 290s over PLX is perfectly OK.
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#7
The 2FPS difference between the 3770k and 3930k is the fact that a 3770k is faster in single threaded tasks and has nothing to do with the PLX chip. Basically the PLX chip will be slower than native PCI-e by XXX nano seconds. In Xfire the PCI-e bus is used for shipping rendered images from the second card to the first and for the first to tell the second when to start working. The moving of a raw 1080p image over PCI-e 3.0 x16 takes 377ns from the second card to the cpu and another 377ns from the CPU to the first GPU. With the PLX you just add the delay that the PLX takes to re route and add it to both the GPU2 to CPU and CPU to GPU1 transfer and you get the effect of the PLX chip on the frame rate. The PLX chip on your mobo takes 126ns to reroute so it adds 252ns to the time between new frames which is so little that you won't ever notice.
So Xfire r9 290s over PLX is perfectly OK.
Thanks. logically yes it should work. I ask Ian of Anandtech, (thanks Ian for the quick reply) and he said it should be perfectly ok.....

I wish there is return policy here in Korea but the fact that once it was opened, no return but I can exchange if there is a problem.

He just dont want to pull the trigger yet cause of this return policy. without a firm evidence that it should work........
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#8
bump? anybody who has more input?
 
Joined
Apr 3, 2012
Messages
3,506 (1.66/day)
Likes
7,429
Location
St. Paul, MN
System Name LappyTop/ Bay2- Lowerbay/Opticon As of 6/21/17
Processor i5 1.7GHz/ i3 2120's/i5 2400
Motherboard ASUStek/ HP mini/Dell
Cooling Air/air/air
Memory 8GB Kingston HyperX 1600/ 2GB/ 4GB samsung
Video Card(s) intel HD4000/ HD2000's/ HD2000
Storage 256GB Crucial Mx100 SSD/ couple of spinners/500GB WD
Display(s) 1366x768/ one acer x223w between the rest
Case ALU, beauty!/ HP SFF/ Dell Mini
Audio Device(s) Sonic Master/ onboard's/ onboard
Power Supply ASUS brick, 17W i5/ 19.5 volt bricks/ Dell 265W
Mouse cheap GigaWire930, CMStorm Havoc and a Logitech M510 wireless
Keyboard eMachines serial and a Dynex usb
Software WIN8.1-64bit/ Mint 18 Sarah/Mint 18
Benchmark Scores 6.9 processes per second!/ cruncher's!!
#9
IDK, but, sounds like you could take GPU 1 and cover it in chocolate, GPU 2 and cover it in Peanut butter. In theory, according to Hoyle, it would only add 114ns to the PLX chip total. Meaning the transfer rate would still be quicker than the time between when you think you are gonna sneeze, and you close your eyes...to sneeze.

Much quicker.

:roll:
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#10
IDK, but, sounds like you could take GPU 1 and cover it in chocolate, GPU 2 and cover it in Peanut butter. In theory, according to Hoyle, it would only add 114ns to the PLX chip total. Meaning the transfer rate would still be quicker than the time between when you think you are gonna sneeze, and you close your eyes...to sneeze.

Much quicker.

:roll:
:roll: of course it is much quicker....

its just I am after if there is a huge difference between running them PCIE3 x 8 (dual) directly connected to CPU.

and a crossfired 290's thru PLX chips
 
Joined
Apr 3, 2012
Messages
3,506 (1.66/day)
Likes
7,429
Location
St. Paul, MN
System Name LappyTop/ Bay2- Lowerbay/Opticon As of 6/21/17
Processor i5 1.7GHz/ i3 2120's/i5 2400
Motherboard ASUStek/ HP mini/Dell
Cooling Air/air/air
Memory 8GB Kingston HyperX 1600/ 2GB/ 4GB samsung
Video Card(s) intel HD4000/ HD2000's/ HD2000
Storage 256GB Crucial Mx100 SSD/ couple of spinners/500GB WD
Display(s) 1366x768/ one acer x223w between the rest
Case ALU, beauty!/ HP SFF/ Dell Mini
Audio Device(s) Sonic Master/ onboard's/ onboard
Power Supply ASUS brick, 17W i5/ 19.5 volt bricks/ Dell 265W
Mouse cheap GigaWire930, CMStorm Havoc and a Logitech M510 wireless
Keyboard eMachines serial and a Dynex usb
Software WIN8.1-64bit/ Mint 18 Sarah/Mint 18
Benchmark Scores 6.9 processes per second!/ cruncher's!!
#11
I know you wanted a second opinion, I just goofed off but, from buildzoid 's description... timings should not be an issue. However, if you are like me, less IS more.

So, in theory, fewer gates, fewer hinges to oil.

Hmm, mayhaps I can think of a few more obscurities to describe something as obviously as I can....

:toast:
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#12
Appreciate it by trying to explain more. There is no need for you to do it. I got the logic. But actual impacts is what I am after for.....

look for example 290's vs 780's..... 290's has more bandwidth than gtx 780's. but, its always like that. NVIDIA cards has always less band width than AMD. and all benches is neck to neck comparison with todays games and apps. Its not like head vs toes difference.

But now there is a PLX involve. would there be a performance impact now that 290's are sharing infos with these PLX? Who knows until somebody has actually do a very intensive testing....

IF only, here in korea, there is a 30 days return, I will test it myself once I got hold with the mobo I like with PLX chip..... But rules here is that once you bought it, no return. I can exchange but with same type not different type.....

So........ before actually pulling trigger on 290's I just want to read some benches....

In that way, I can decide if its a good idea to get 290's xfire or tri fired or even quad fire with PLX chips mobos
 
Joined
Apr 3, 2012
Messages
3,506 (1.66/day)
Likes
7,429
Location
St. Paul, MN
System Name LappyTop/ Bay2- Lowerbay/Opticon As of 6/21/17
Processor i5 1.7GHz/ i3 2120's/i5 2400
Motherboard ASUStek/ HP mini/Dell
Cooling Air/air/air
Memory 8GB Kingston HyperX 1600/ 2GB/ 4GB samsung
Video Card(s) intel HD4000/ HD2000's/ HD2000
Storage 256GB Crucial Mx100 SSD/ couple of spinners/500GB WD
Display(s) 1366x768/ one acer x223w between the rest
Case ALU, beauty!/ HP SFF/ Dell Mini
Audio Device(s) Sonic Master/ onboard's/ onboard
Power Supply ASUS brick, 17W i5/ 19.5 volt bricks/ Dell 265W
Mouse cheap GigaWire930, CMStorm Havoc and a Logitech M510 wireless
Keyboard eMachines serial and a Dynex usb
Software WIN8.1-64bit/ Mint 18 Sarah/Mint 18
Benchmark Scores 6.9 processes per second!/ cruncher's!!
#13
In my opinion, as a non tech, is that it unlikely you will have a major meltdown running on a PLX. It may not be the ultimate bandwidth setup, however, it does appear, if what buildzoid 's statements are true, that there will be a minor nanoseconds difference and will be a nice upgrade, despite the blower fan noise.

Perhaps, YOU will be the one doing an extensive review of the 290's xfired on a PLX? ;)
 
Joined
Dec 23, 2012
Messages
1,396 (0.75/day)
Likes
942
Location
south korea
System Name budget gaming / MSI gs70 6qe silver
Processor i7 4770k 4.0 ghz / i7 6700hq
Motherboard Asus Z87m-plus / msi h97
Cooling Zigmatek Janus LP / stock
Memory G.Skill TridentX 16gb 2666 mhz CL11 / Samsung ddr4 2133
Video Card(s) Zotac GTX 1060 6gb AMP / gtx 970m
Storage 750gb Seagate (OS); 1x 1tb WD; 1x2tb Seagate
Display(s) 2x crossover 30x-p driver + crossover 3020mdp 1600p eyefinity / HP elitedisplay E231 1080p
Case BRAVOTEC DEFY R50
Audio Device(s) Onboard / onboard
Power Supply Silverstone 750W gold/ Powertex 500w
Mouse Logitech G100S
Keyboard Steelseries Mechanical keyboard
Software Windows 10
#14
well I can but I dont have 780ti's to compare or other cards to compare :p

Paging AMD and NVIDIA please lend me 2 or 3 290x and 2 o3 780ti's so I can have extensive review about these gpu's with PLX involve

:clap::peace::lovetpu: