- Mar 15, 2014
- 78
- 0
- 0
There are lots of Mantle Reviews on the net but none of them quite seem to reflect the point of Mantle and what it does.
My experience of those reviews seems typical in that they have a tendancy to use the best 12 thread Intel CPU's overclocked and then set very high in game settings, often also only testing in Campaign mode which has far less CPU implications.
What this does is make the Game completely GPU bound.
Mantle removes or betters CPU bottlenecking, so.
Anyone worth their knowledge knows the way you test for CPU bettlenecks is to run very low GFX setting, this moves the workload away from the GPU onto the CPU, this is what i have seen Anand do many times, and its right.
So with that principle in mind i have done some testing of my own, little of it, but didn't need to to show its effects right away.
So, here we go.
System used:
Motherboard: Asus Sabertooth 990FX, Rev-1.01, BIOS 1503.
CPU: AMD Phenom II x6 1090T @ 4.2Ghz
GPU: Radeon 7870XT (Tahiti LE) @ 1150 / 1575
RAM: Corsair XMS3 4x2GB @ 1600Mhz
SSD Samsung 830
PSU: XFX 750 Watt Pro Edition
Case: CM HAF 912+
Op: Shanghai:
I chose this map immediately as it has a well known CPU bottleneck; all the way down the main dual bridge road. Standing at the very top of the road my results are as follows.
48 Player server, with 40+ players active. (GFX setting = Medium preset @ 1080P)
In Direct X: 57 FPS.
In Mantle: 123 FPS
The performance difference in Mantle is +110% in that scenario.
This does not mean there is a FPS boost everywhere all the time, thats not the case at all.
I play BF4 all the time, since Mantle became available in BF4 i have switched to it and never looked back.
In Direct X my FPS get dragged down by a half regularly, often less than what you see in that screenshot.
With Mantle i'm constantly running roughly around 100 FPS Minimum on all High Settings 1080P. its like night and day.
I have made a couple of Youtube vids to show it in action, keep in mind that what i used to record it, Mirillis Action, reduces the performance by a good 20%.
Direct X
https://www.youtube.com/watch?v=3WFqX5k_GgA
Mantle
https://www.youtube.com/watch?v=pcEEuIvN8FI
4.8Ghz Sandy Bridge i7 and R9 290 xFire x2 test in this post. http://forums.anandtech.com/showpost.php?p=36169582&postcount=40
My experience of those reviews seems typical in that they have a tendancy to use the best 12 thread Intel CPU's overclocked and then set very high in game settings, often also only testing in Campaign mode which has far less CPU implications.
What this does is make the Game completely GPU bound.
Mantle removes or betters CPU bottlenecking, so.
Anyone worth their knowledge knows the way you test for CPU bettlenecks is to run very low GFX setting, this moves the workload away from the GPU onto the CPU, this is what i have seen Anand do many times, and its right.
So with that principle in mind i have done some testing of my own, little of it, but didn't need to to show its effects right away.
So, here we go.
System used:
Motherboard: Asus Sabertooth 990FX, Rev-1.01, BIOS 1503.
CPU: AMD Phenom II x6 1090T @ 4.2Ghz
GPU: Radeon 7870XT (Tahiti LE) @ 1150 / 1575
RAM: Corsair XMS3 4x2GB @ 1600Mhz
SSD Samsung 830
PSU: XFX 750 Watt Pro Edition
Case: CM HAF 912+
Op: Shanghai:
I chose this map immediately as it has a well known CPU bottleneck; all the way down the main dual bridge road. Standing at the very top of the road my results are as follows.
48 Player server, with 40+ players active. (GFX setting = Medium preset @ 1080P)
In Direct X: 57 FPS.
In Mantle: 123 FPS
The performance difference in Mantle is +110% in that scenario.
This does not mean there is a FPS boost everywhere all the time, thats not the case at all.
I play BF4 all the time, since Mantle became available in BF4 i have switched to it and never looked back.
In Direct X my FPS get dragged down by a half regularly, often less than what you see in that screenshot.
With Mantle i'm constantly running roughly around 100 FPS Minimum on all High Settings 1080P. its like night and day.
I have made a couple of Youtube vids to show it in action, keep in mind that what i used to record it, Mirillis Action, reduces the performance by a good 20%.
Direct X
https://www.youtube.com/watch?v=3WFqX5k_GgA
Mantle
https://www.youtube.com/watch?v=pcEEuIvN8FI
4.8Ghz Sandy Bridge i7 and R9 290 xFire x2 test in this post. http://forums.anandtech.com/showpost.php?p=36169582&postcount=40
Last edited: