• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

R9 Nano Review thread

humbug, your combined gpu/physics test score is lower. it's not the sum of the separate tests that's the problem, it's the last combined test thats test both simultaneously. for whatever reason, it's not running as well on AMD cpus.

both my GPU and CPU scores are higher, so explain to me how higher scores result in lower scores?
 
Last edited:
Just bought one for a x99/5820k/512GB SM915 mitx setup.
Just built some full atx rigs with same spec except using GTX 980 instead.
Would be interesting to see them compare.
Will try it when i get chance sometime.
What's everything using to benchtest their full system nowadays?
 
both my GPU and CPU cores are higher, so explain to my how higher scores result in lower scores?

ok:)


There are 4 tests. 2x gpu tests, 1x physics and 1x combined gpu+physics.

- yours is faster in the gpu and physics tests, as it should be you have a highly clocked fx vs my stock i5 and a higher clocked GPU.
- mine is faster in the combined test. Dont know why.

The overall score is, clearly, reliant on the combined test. Therefor, despite mine being slower in the separate gpu and physics tests, being faster in the combined test tips the score in my favor. And mine is, percentage-wise, 50% faster or something like that.

So, obviously something is up with the 3rd test on AMD, possible only the FX line, not sure.
 
Last edited:
ok:)


There are 3 test groups, 1) gpu (which consists of two separate tests) 2)physics and 3)gpu+physics.

- yours is faster in group one and 2, as it should be you have a highly clocked fx and a higher clocked GPU.
- mine is faster in the 3rd. Dont know why.

The overall score is, clearly, reliant on the 3rd combined test. Therefor, despite mine being slower in group 1 and 2, being faster in test 3 tips the score in my favor. And mine is, percentage-wise, 50% faster or something like that.

So, obviously something is up with the 3rd test on AMD, possible only the FX line, not sure.

I have asked them, if the performance is X in Physics there is no reason why the performance should be half on another run on the same engine.

It looks like they are using 8 threads on the Physics run but only 4 on the Physics + Graphics run, and thats BS..... if the engine is capable on running 8 threads for Physics then it should use 8 threads, it does not differentiate with Intel, as it shouldn't, nor should it with AMD.
 
Last edited:
It's pretty low end to pair with a high end card is all. It's going to hold you back judging by the physics score you got. For instance my old shed of a cpu scores 9900 in the Physics test. In fact my rig scores similar to yours with my card at stock and is not even worth what your Nano cost.

Edit: if you were capped then fair enough. I think i would still want an i7 quad core in there to make sure i am getting the best from the nano.

At stock nano should be getting close to 14k graphics score according to this.

http://www.pcper.com/reviews/Graphics-Cards/AMD-Radeon-R9-Nano-Review/3DMark-Fire-Strike-and-Unigine-Heaven

I have asked them, if the performance is X in Physics there is no reason why the performance should be half on another run on the same engine.

It looks like they are using 8 threads on the Physics run but only 4 on the Physics + Graphics run, and thats BS..... if the engine is capable on running 8 threads for Physics then it should use 8 threads, it does not differentiate with Intel, as it shouldn't, nor should it with AMD.

Should be able to test that with afterburner graphs or even task manager.
 
I have asked them, if the performance is X in Physics there is no reason why the performance should be half on another run on the same engine.

It looks like they are using 8 threads on the Physics run but only 4 on the Physics + Graphics run, and thats BS..... if the engine is capable on running 8 threads for Physics then it should use 8 threads, it does not differentiate with Intel, as it shouldn't, nor should it with AMD.

It is not just you. All the results using The FX-9590 have low combined scores.... \

I totally agree with you that your score should be considerably higher as your GFX and Physics are fine.....

Here is a 31k!!!!!!! graphics score using the FX-9590

something is definitely up..... I just cant see how that works.....
 
http://s3.amazonaws.com/download-aws.futuremark.com/3DMark_Technical_Guide.pdf

Physics test

3DMark Fire Strike Physics test benchmarks the hardware’s ability to run gameplay physics simulations on the CPU. The GPU load is kept as low as possible to ensure that only the CPU
is stressed. The Bullet Open Source Physics Library is used as the physics library for the test.

The test has 32 simulated worlds. One thread per available CPU core is used to run simulations. All physics are computed on CPU with soft body vertex data updated to GPU
each frame.

Combined test

3DMark Fire Strike Combined test stresses both the GPU and CPU simultaneously. The GPU load combines elements from Graphics test 1 and 2 using tessellation, volumetric illumination, fluid simulation, particle simulation, FFT based bloom and depth of field.

The CPU load comes from the rigid body physics of the breaking statues in the background. There are 32 simulation worlds running in separate threads each containing one statue decomposing into 113 parts. Additionally there are 16 invisible rigid bodies in each world except the one closest to camera to push the decomposed elements apart. The simulations run on one thread per available CPU core.

The 3DMark Fire Strike Combined test uses the Bullet Open Source Physics Library

Multithreading

The multithreading model is based on DX11 deferred device contexts and command lists.

The engine utilizes one thread per available CPU core. One of the threads is considered as the main thread, which uses both immediate device context and deferred device context.
The other threads are worker threads, which use only deferred device contexts.

Rendering workload is distributed between the threads by distributing items (e.g. geometries and lights) in the rendered scene to the threads. Each thread is assigned roughly equal
amount of scene items. When rendering a frame, each thread does the work associated to items assigned to the
thread. That includes, for example, computation of transformation matrix hierarchies, computation of shader parameters (constants buffer contents and dynamic vertex data) and
recording of DX API calls to a command list. When the main thread is finished with the tasks associated to its own items, it executes the command lists recorded by worker threads.

so it should be using 8 threads, but perhaps the shared fpus are holding the performance back in the combined test when all cores are loaded?
 
Last edited:
It is not just you. All the results using The FX-9590 have low combined scores.... \

I totally agree with you that your score should be considerably higher as your GFX and Physics are fine.....

Here is a 31k!!!!!!! graphics score using the FX-9590

something is definitely up..... I just cant see how that works.....

Yup.

Its no secrete AMD CPU's are a bit pap for DX9/10/11 Draw Calls as low threading performance is not much good compared with Intel.

But there is far more to gaming than Draw Calls, AMD's 8 core Vishera are rather good at Multi-Threaded Physics, better even than older gen 8 thread i7's, certainly Sandy Bridge and at least on par with Ivy Bridge, a CPU of the same era.

If the CPU Physics hold back the GPU in the Physics test, which they do, then the Physics will do the same in the combined test, which again they do.
It should be to the same extent as the Graphics being bottlenecked play no part in the render performance, it is capped by the Physics performance and they cannot be capped to a different extent on different parts if that capping is dictated to differ or not depending on what brand of CPU it is.

Its a crock of _____
 
Last edited:
http://s3.amazonaws.com/download-aws.futuremark.com/3DMark_Technical_Guide.pdf







so it should be using 8 threads, but perhaps the shared fpus are holding the performance back in the combined test when all cores are loaded?

Only if the engine is using the FPU's incorrectly, bad coding.

Vishara is a programmable modular design, it does differ in that way from Intel which is fixed.

To explain.

Take an Intel 4 core 8 thread.
4 Integer units with 2 FP threads each
or 4 Integer 4 FP for the i5

That's what it is and that's what you get ^^^^

AMD's 8 core Vishera looks like this.

4x 2 Integer in a split 256Bit FP Module.

So you have 4 modules with two 128Bit FP threads sharing one L2 Cache

If desired you can tell one, two, three or all Modules to comine the two 128Bit threads from the two Integer units to make one big integer unit / 256Bit FP thread

The idea being if you have lots of little things going on but on big thing you can have 6 threads all with their own calc and float units and one big fat one to do the high workload stuff... or a combination of whatever that ^^^^

Its actually a very clever architecture and works well, if you have a brain to use it as intended.

If not you can end up with it configured wrongly for the task and it perform like crap.
 
Well done for posting the anti frtc up, the g score of 14000 is good.
3d mark is irrelevant to me, the only relevant score in the bench is the graphics score, even then it's still irrelevant :)

Yes That's much better :)
 
Last edited:
Yeah but we knew that. And now we know the performance of the Fury X2 or there abouts. It will be the king of single card gaming, but at what cost? If they can do it for under £1000 I'll be impressed.
 
Back
Top Bottom