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

Rise of The Tomb Raider DX12

Joined
Feb 13, 2016
Messages
37 (0.05/day)
Likes
21
Location
Buenos Aires, Argentina
System Name Monster
Processor FX 8320 oc to 3.6Ghz
Motherboard Gigabyte 97A-UD3P
Cooling Corsair H100i
Memory Kingston 16Gb
Video Card(s) Asus Strix RX580 OC 8Gb
Storage 250Gb Samsung EVO SSD plus three spinners up to 1Tb
Display(s) Samsung C27F591 curved monitor 27" with FreeSync
Case Cooler Master Storm Trooper
Audio Device(s) VIA crap
Power Supply Sentey 850W
Mouse AFX LMO216
Keyboard Corsair Strafe RGB Silent
Software Windows 10 x64
Benchmark Scores 3DMark 4053 (lower with freesync off)
#1
There's a lot of talk about DX12 being sluggish in this game vs DX11, so I ran the in-game benchmark and found the DX12 performance to be nearly 20fps better than DX11.
Not sure if this is a measure of how well the game runs in DX12, but for me it seems to perform according to the benchmark IE about 20fps better. The only other DX12 games I have are the Forza racing games (7, Apex and Horizon) but it's bit hard to compare since you can't switch to DX11 for those.
rottr-dx12-dx11-benchmark.jpg
 
Joined
Mar 13, 2009
Messages
822 (0.25/day)
Likes
98
Location
Bannock County
Processor AMD FX-8320 4.1Ghz
Motherboard Asus Aura
Cooling Corsair H60
Memory Crucial Ballistix Sport 32GB
Video Card(s) EVGA 1060 SC
Storage Intel 535 Series SSD + Seagate FireCuda 2TB
Case NZXT
Power Supply EVGA 650 Watt
Software Windows 10 Pro 64-Bit
#2
Yep, I gain about 15fps in directx 12 with this game.
 
Joined
Nov 9, 2010
Messages
5,194 (1.95/day)
Likes
1,734
Processor Intel i7 950 @ 3.2GHz
Motherboard ASUS P6X58D-E
Cooling Corsair H50 push/pull
Memory Kingston HyperX 1600 8GB
Video Card(s) Sapphire HD 7970 OC
Storage Plextor M5P 128GB/WD Black 2x1TB,1x6TB/Seagate 1TB
Display(s) Panasonic TC-L32U3
Case Antec DF-85
Audio Device(s) Yamaha RX-V371 AVR
Power Supply XFX 850w Black Edition
Mouse Logitech G402
Keyboard Logitech K120
Software W10 Pro 64 bit
#3
How much difference in performance you get between Dx11 and Dx12 depends on the game and the GPU. The more recent titles and latest AMD GPUs are going to yield best case scenario. I however on my 7970 with fewer ACE units don't see much difference.
 
Joined
Oct 2, 2015
Messages
836 (0.95/day)
Likes
379
Location
Argentina
System Name Ciel / Yukino
Processor AMD Ryzen R3 1200 @ 3.85GHz / Intel Core i3 5005U @ 2GHz
Motherboard MSI B350M PRO-VDH / HP 240 G5
Cooling Wraith Stealth / Stock
Memory 2x 4GB Corsair Ballistix Sport DDR4 2400MHz @ 3000MHz CL16 / 2x 4GB Hynix + Kingston DDR3L 1600MHz
Video Card(s) XFX AMD HD 7750 1GB @ 900/1250MHz / Intel HD 5500
Storage HDD Hitachi 1TB / SSD Kingston 32GB
Display(s) HP w17e 1440x900 @ 75 Hz / Integrated 1366x768 @ 94Hz
Case Generic / Stock
Audio Device(s) Realtek ALC892 / Realtek ALC282
Power Supply Sentry XPP 525W / Power Brick
Mouse Noganet ST-334 2400DPI @ 1000Hz / Touchpad
Keyboard Generic PS2 / Stock
Software Windows 10 LTSB 2016 x64 + Gentoo x64
Benchmark Scores Fire Strike: 2687 Time Spy: 546
#4
Both DX12 and Vulkan are a god send for weak CPU users, I remember trying Battlefield 1 in an A4-4000 (a single module FX based APU), DX11 got 15FPS, DX12 got 40.
You won't see a huge difference with an i7 and a 1080ti, but on cheaper hardware, it helps a lot.
 

cadaveca

My name is Dave
Staff member
Joined
Apr 10, 2006
Messages
16,601 (3.83/day)
Likes
11,013
Location
Parkland County, Alberta
#5
How much difference in performance you get between Dx11 and Dx12 depends on the game and the GPU. The more recent titles and latest AMD GPUs are going to yield best case scenario. I however on my 7970 with fewer ACE units don't see much difference.
No. It depends more on the CPU. That is why the OP sees such a large increase... they've got a AMD FX-8350.

Both DX12 and Vulkan are a god send for weak CPU users, I remember trying Battlefield 1 in an A4-4000 (a single module FX based APU), DX11 got 15FPS, DX12 got 40.
You won't see a huge difference with an i7 and a 1080ti, but on cheaper hardware, it helps a lot.
Exactly.


There's a lot of talk about DX12 being sluggish in this game vs DX11, so I ran the in-game benchmark and found the DX12 performance to be nearly 20fps better than DX11.
Not sure if this is a measure of how well the game runs in DX12, but for me it seems to perform according to the benchmark IE about 20fps better. The only other DX12 games I have are the Forza racing games (7, Apex and Horizon) but it's bit hard to compare since you can't switch to DX11 for those.
View attachment 96051
Your CPU is rather weak compared to more recent releases, so you should see a good benefit from a lot of titles that offer DX12.
 
Joined
Dec 16, 2012
Messages
1,733 (0.91/day)
Likes
1,414
Location
Finland
System Name SER
Processor Intel Core i5-7600K @ 4.7GHz
Motherboard Asus Z170 Pro Gaming
Cooling Alphacool Eisbaer 240 + 4x Corsair ML120
Memory 16GB Corsair Vengeance LPX DDR4-2400 @ 2666MHz
Video Card(s) Asus GeForce GTX 980 Strix @ 1500/2000MHz
Storage 2x 256GB, 2x 250GB, 240GB & 120GB SSD
Display(s) 2x 1920x1080
Case Phanteks Eclipse P400S Windowed
Audio Device(s) Creative Sound Blaster Z
Power Supply Super Flower Golden Green 500W
Mouse Logitech G400s
Keyboard Logitech G410
Software Windows 10 Pro EN
#6
It runs slightly better on my PC with DX12 than DX11. i5-7600K @ 4.7GHz, 16GB @ 2666MHz, GTX 980 @ 1500/2000.
 
Joined
Nov 9, 2010
Messages
5,194 (1.95/day)
Likes
1,734
Processor Intel i7 950 @ 3.2GHz
Motherboard ASUS P6X58D-E
Cooling Corsair H50 push/pull
Memory Kingston HyperX 1600 8GB
Video Card(s) Sapphire HD 7970 OC
Storage Plextor M5P 128GB/WD Black 2x1TB,1x6TB/Seagate 1TB
Display(s) Panasonic TC-L32U3
Case Antec DF-85
Audio Device(s) Yamaha RX-V371 AVR
Power Supply XFX 850w Black Edition
Mouse Logitech G402
Keyboard Logitech K120
Software W10 Pro 64 bit
#7
No. It depends more on the CPU. That is why the OP sees such a large increase... they've got a AMD FX-8350.
Sounds like you're unaware of what ACE units are on an AMD GPU and what they do. They're necessary for proper Dx12 support, and the older AMD GPUs are very anemic when it comes to ACE.
 
Joined
Aug 20, 2007
Messages
8,200 (2.13/day)
Likes
7,216
System Name New Genesis
Processor AMD Ryzen 1800X @ 4.1Ghz All Cores
Motherboard GIGABYTE X370 Aorus Gaming 5
Cooling A whole lotta Sunon and Corsair Maglev blower fans...
Memory G.SKILL TridentZ Series 32GB (4 x 8GB) DDR4-3200 @ 14-14-14-34-1T
Video Card(s) NVIDIA Titan XP Star Wars Collectors Edition (Galactic Empire)
Storage HGST Ultrastar 7k6000 2TB HDD (128MBs of Cache)
Display(s) BenQ BL3200PT (a 1440p VA Panel with decent latency)
Case Thermaltake Core X31
Audio Device(s) Onboard Toslink to Schiit Modi Multibit to Asgard 2 Amp to AKG K7XX Ruby Red Massdrop Headphones
Power Supply Seasonic PRIME 750W 80Plus Titanium
Mouse ROCCAT Kone EMP
Keyboard WASD CODE 104-Key w/ Cherry MX Green Keyswitches, Doubleshot Vortex PBT White Keycaps, Blue legends
Software Windows 10 Enterprise (From former workplace, yay no telemetry)
Benchmark Scores FS/TS Extreme FS 14011:https://www.3dmark.com/fs/14938447 TS 4769:https://www.3dmark.com/spy/3315715
#8
Sounds like you're unaware of what ACE units are on an AMD GPU and what they do. They're necessary for proper Dx12 support, and the older AMD GPUs are very anemic when it comes to ACE.
And DX12 lowers overhead where, exactly?

CPU.
 

cadaveca

My name is Dave
Staff member
Joined
Apr 10, 2006
Messages
16,601 (3.83/day)
Likes
11,013
Location
Parkland County, Alberta
#9
Sounds like you're unaware of what ACE units are on an AMD GPU and what they do. They're necessary for proper Dx12 support, and the older AMD GPUs are very anemic when it comes to ACE.
Well, to be honest, you are correct that different GPUs will see different gains, but again, as RTB posted, it's the CPU overhead that is changed by DX12. You have a very well-balanced system, so you should not see much in the way of gains really, but those that have an under-powered CPU for their GPU will see far better gains with the same card you have now. BTW, I do still have a 7970, and did do some DX12 testing with it just like the OP a couple of weeks ago, investigating this exact thing.


BTW, ACE = Asynchronous Compute Engine. I actually am pretty aware of what they are, and what they do, and why an "ACE" is important.

Anyway, the OP has an RX580; it's not an older GPU. So yeah, your comment does have some merit, but does not apply to the OP's results really. That's why I responded in the way I did.. I looked at the OP's system config and based my coment on his experience with his specific hardware

As to why I say what I do, rather than just say stuff, let's go to those that make the cards, and see what they say about what DX12 offers, and how it can boost performance...

Here's what AMD says about DX12:

DirectX 12 takes your game to a whole new level. By tapping into all the cores of your multicore CPU simultaneously, DirectX 12 boosts the amount of work that can be sent to the graphics card – and gets it done in less time. Plus, DirectX 12 allows your GPU to run multiple tasks at once, so you save rendering time, reduce latency, and get more frames per second. Two revolutionary features in DirectX 12 are driving these innovations: multi-threaded command buffer recording and asynchronous shaders.
https://www.amd.com/en/technologies/directx12

Nvidia says:

In addition to new and more realistic visual effects, DirectX 12 promises to improve performance through a number of API changes, software features, and hardware features. Of these improvements, draw call optimisations are likely to have the most immediate impact, relieving CPU bottlenecks in open world games, and in large-scale strategy titles.
https://www.geforce.co.uk/whats-new/articles/geforce-gtx-is-game-ready-for-windows-10-and-directx-12
 
Last edited:
Joined
Jan 8, 2017
Messages
1,981 (4.81/day)
Likes
1,136
System Name Good enough
Processor AMD FX-6300 - 4.5 Ghz
Motherboard ASRock 970M Pro3
Cooling Scythe Katana 4 - 3x 120mm case fans
Memory 16GB - 4x4GB A-DATA 1866 Mhz (OC)
Video Card(s) ASUS GTX 1060 Turbo 6GB ~ 2139 Mhz / 9.4 Gbps
Storage 1x Samsung 850 EVO 250GB , 1x 1 Tb Seagate something or other
Display(s) 1080p TV
Case Zalman R1
Power Supply 500W
#10
And DX12 lowers overhead where, exactly?

CPU.
What is baffling though is why there is a regression in performance on higher end CPUs.
 
Joined
Nov 9, 2010
Messages
5,194 (1.95/day)
Likes
1,734
Processor Intel i7 950 @ 3.2GHz
Motherboard ASUS P6X58D-E
Cooling Corsair H50 push/pull
Memory Kingston HyperX 1600 8GB
Video Card(s) Sapphire HD 7970 OC
Storage Plextor M5P 128GB/WD Black 2x1TB,1x6TB/Seagate 1TB
Display(s) Panasonic TC-L32U3
Case Antec DF-85
Audio Device(s) Yamaha RX-V371 AVR
Power Supply XFX 850w Black Edition
Mouse Logitech G402
Keyboard Logitech K120
Software W10 Pro 64 bit
#11
Well, to be honest, you are correct that different GPUs will see different gains, but again, as RTB posted, it's the CPU overhead that is changed by DX12.
I feel it's kind of misleading to use the worst case scenario to measure gains. Are we really talking about gains there, or just what a common sense rig can do verses a poorly built one? It's the same as saying Ryzen is way better than Bulldozer, or that AMD makes better use of Dx12 than Nvidia, without noting that AMD is still struggling to keep up with Nvidia in actual performance.

The bottom line is, components already achieving great performance have a lower ceiling for things like OCing and Dx12.
 
Joined
Feb 13, 2016
Messages
37 (0.05/day)
Likes
21
Location
Buenos Aires, Argentina
System Name Monster
Processor FX 8320 oc to 3.6Ghz
Motherboard Gigabyte 97A-UD3P
Cooling Corsair H100i
Memory Kingston 16Gb
Video Card(s) Asus Strix RX580 OC 8Gb
Storage 250Gb Samsung EVO SSD plus three spinners up to 1Tb
Display(s) Samsung C27F591 curved monitor 27" with FreeSync
Case Cooler Master Storm Trooper
Audio Device(s) VIA crap
Power Supply Sentey 850W
Mouse AFX LMO216
Keyboard Corsair Strafe RGB Silent
Software Windows 10 x64
Benchmark Scores 3DMark 4053 (lower with freesync off)
#12
That explains a lot, so thanks. I've had the FX8320 for three or four years now and only recently switched from a GTX970 to an RX580 and if my memory serves me correctly, DX12 performance on the 970 wasn't as good as now.
I must have missed ACE when I was looking to change the GPU, so this comes as an extra bonus and the game with the most noticeable improvement is Forza Horizon 3, which is a DX12 game.
 
Joined
Jan 8, 2017
Messages
1,981 (4.81/day)
Likes
1,136
System Name Good enough
Processor AMD FX-6300 - 4.5 Ghz
Motherboard ASRock 970M Pro3
Cooling Scythe Katana 4 - 3x 120mm case fans
Memory 16GB - 4x4GB A-DATA 1866 Mhz (OC)
Video Card(s) ASUS GTX 1060 Turbo 6GB ~ 2139 Mhz / 9.4 Gbps
Storage 1x Samsung 850 EVO 250GB , 1x 1 Tb Seagate something or other
Display(s) 1080p TV
Case Zalman R1
Power Supply 500W
#13
ACE is not causing lower CPU overhead.

ACE gives GPUs the capability to switch contex without a big performance penalty. It's useful when you want to do compute shaders at the same time as graphics. This is why for example Doom sees a bigger uplift on Vulkan on AMD cards , that's a game that makes extensive use of GPU accelerated particles.

In this case though the performance improvement comes pretty much exclusively from the multithreaded draw calls.
 
Joined
Nov 9, 2010
Messages
5,194 (1.95/day)
Likes
1,734
Processor Intel i7 950 @ 3.2GHz
Motherboard ASUS P6X58D-E
Cooling Corsair H50 push/pull
Memory Kingston HyperX 1600 8GB
Video Card(s) Sapphire HD 7970 OC
Storage Plextor M5P 128GB/WD Black 2x1TB,1x6TB/Seagate 1TB
Display(s) Panasonic TC-L32U3
Case Antec DF-85
Audio Device(s) Yamaha RX-V371 AVR
Power Supply XFX 850w Black Edition
Mouse Logitech G402
Keyboard Logitech K120
Software W10 Pro 64 bit
#14
That explains a lot, so thanks. I've had the FX8320 for three or four years now and only recently switched from a GTX970 to an RX580 and if my memory serves me correctly, DX12 performance on the 970 wasn't as good as now.
I must have missed ACE when I was looking to change the GPU, so this comes as an extra bonus and the game with the most noticeable improvement is Forza Horizon 3, which is a DX12 game.
Yeah but the 970 has less than half the VRAM of the 580 too. You're always going to notice more FPS from Dx11 to Dx12 on a AMD GPU because they use more compute power specifically designed for Dx12, but their higher end models still can't keep up with Nvidia's ones.

Also, despite the 970 and 580 being close in release price, they are not close in launch date. As time goes on things like high VRAM capacity and more processing power become more common and more affordable, so you really can't compare the two apples to apples. The current Nvidia GPU to compare to the 580 on price and performance is the 1060 6GB.

Personally I'll take less improvement on Dx for way more performance to begin with. If you don't look at it in that regard, it's very misleading. It's the GPUs with better performance to begin with that are going to have less gain, just like factory overclocked GPUs have less headroom for user overclocking.
 
Last edited:

lyndonguitar

I play games
Joined
Apr 1, 2010
Messages
1,872 (0.65/day)
Likes
530
Location
Philippines
System Name X6 | Lyndon-ROG
Processor Intel Core i7-8700k | Intel Core i7 6700HQ
Motherboard Gigabyte Z370 Aorus Gaming 5 | Asus ROG-GL552VX
Cooling Deepcool Captain 240EX
Memory 16GB Corsair Vengeance LED | 8 GB
Video Card(s) NVIDIA GTX 1080 8 GB GDDR5X | NVIDIA GeForce GTX 950M 4GB
Storage SSDs: 500GB, HDDs: 2TB, 2TB, 3TB | SSD: 250GB, HDD: 1TB
Display(s) BenQ XL2411 1080p@144Hz, 32" HDTV, 19" Monitor | 15.6"1080p
Case Cougar Panzer Max
Audio Device(s) Corsair Gaming H1500 7.1 | ROCCAT Kave 5.1 | Edifier M3200
Power Supply EVGA 750GQ
Mouse Razer Deathadder Chroma | Logitech G302 | Mad Catz Cyborg R.A.T. 5
Keyboard Corsair Vengeance K70 Cherry MX Red
Software Windows 10
#15
theres no difference on my 8700k and 1080. however I prefer DX11 since it is smoother when playing on Steam Link. DX12 has slight stutters