Overview
Putting your APU to work for more than just graphics
It's clear by now that AMD's latest CPU releases, the Ryzen 3 2200G and the Ryzen 5 2400G are compelling products. We've already taken a look at them in our initial review, as well as investigated how memory speed affected the graphics performance of the internal GPU but it seemed there was something missing.
Recently, it's been painfully clear that GPUs excel at more than just graphics rendering. With the rise of cryptocurrency mining, OpenCL and CUDA performance are as important as ever.
Cryptocurrency mining certainly isn't the only application where having a powerful GPU can help system performance. We set out to see how much of an advantage the Radeon Vega 11 graphics in the Ryzen 5 2400G provided over the significantly less powerful UHD 630 graphics in the Intel i5-8400.
Test System Setup | |
CPU | AMD Ryzen 5 2400G Intel Core i5-8400 |
Motherboard | Gigabyte AB350N-Gaming WiFi ASUS STRIX Z370-E Gaming |
Memory | 2 x 8GB G.SKILL FlareX DDR4-3200 (All memory running at 3200 MHz) |
Storage | Corsair Neutron XTi 480 SSD |
Sound Card | On-board |
Graphics Card | AMD Radeon Vega 11 Graphics Intel UHD 630 Graphics |
Graphics Drivers | AMD 17.40.3701 Intel 23.20.16.4901 |
Power Supply | Corsair RM1000x |
Operating System | Windows 10 Pro x64 RS3 |
GPGPU Compute
Before we take a look at some real-world examples of where a powerful GPU can be utilized, let's look at the relative power of the Vega 11 graphics on the Ryzen 5 2400G compared to the UHD 630 graphics on the Intel i5-8400.
SiSoft Sandra is a suite of benchmarks covering a wide array of system hardware and functionality, including an extensive range of GPGPU tests, which we are looking at today.
Comparing the raw shader performance of the Ryzen 5 2400G and the Intel i5-8400 provides a clear snapshot of what we are dealing with. In every precision category, the Vega 11 graphics in the AMD part are significantly more powerful than the Intel UHD 630 graphics. This all combines to provide a 175% increase in aggregate shader performance over Intel for the AMD part.
Now that we've taken a look at the theoretical power of these GPUs, let's see how they perform in real-world applications.
Video Transcoding
Generally, we test integrated GPU transcoding performance for Intel's CPUs using Handbrake. However, Handbrake has yet to support AMD's APP SDK to enable hardware-accelerated video transcoding.
Luckily, we found an alternative in DVDFab, a popular application for ripping and converting formats such as DVD and Blu-Ray. For our test, we are using a 1080p Blu-Ray ISO source, and transcoding it to a smaller filesize 1080p H264 version with an average bitrate of 5Mbps.
In DVDFab, it's clear that there is an advantage to encoding video on the GPU instead of purely on the CPU cores. Both the AMD and Intel parts see much shorter video transcode times when running on their GPUs, with the Ryzen 5 2400G seeing a 60% speedup. While there can be a quality difference between GPU video encoders versus CPU encoders, the resulting files were very similar in our visual inspection in this particular case.
Despite the advantages in raw compute power, the AMD Ryzen 5 2500G is significantly slower than the Intel Quick Sync-enabled i5-8400. The answer to this most likely lies in optimization.
While Quick Sync has been updated frequently in cadence with new processor releases, the last update for AMP's APP SDK was nearly 2 years ago. In fact, APP has been superseded by a replacement framework, AMD's Advanced Media Framework (AMF). However, we have been unable to find any software support for this toolset beyond AMD's own ReLive application.
We'd love to see AMF support added to an application like Handbrake to allow owners of devices with modern AMD graphics to be able to fully utilize the video encoding features of their GPUs.
Rendering
Cinebench R15
The performance depends on various factors, such as the GPU processor on your hardware, on the drivers used. The graphics card has to display a huge amount of geometry (nearly 1 million polygons) and textures, as well as a variety of effects, such as environments, bump maps, transparency, lighting and more to evaluate the performance across different disciplines and give a good average overview of the capabilities of your graphics hardware. The result is measured in frames per second (fps). The higher the number, the faster your graphics card is.
A go-to CPU benchmark, Cinebench R15 also contains an OpenGL-based test meant to stress the rendering power of GPUs both integrated and discrete.
OpenGL performance of both of these parts is similar, with the Ryzen 5 2400G sporting a 5% performance advantage over the Intel i5-8400.
LuxMark
OpenCL rendering performance is important for workstation-level graphics card hardware. Luxmark, one of the most widely used OpenCL performance tests, provides a good look at how different GPUs perform in typical OpenCL rendering workloads.
For this test, we are using the "Neumann" scene in Luxmark 3.1—the middle compute intensity scene, consisting of almost 1,769,000 triangles.
While pure CPU-based rendering on the Ryzen 5 2400G and the Core i5-8400 are very similar performance wise, moving to GPU rendering on the AMD processor provides an almost 40% performance advantage over the Intel CPU.
Interestingly, the Intel i5-8400 shows very little scaling running on the CPU versus the GPU in OpenCL rendering.
V-Ray Benchmark 1.0.7
V-Ray is popular third-party renderer that plugs into the most powerful CAD and 3D modeling applications. With plugins for 3ds Max, Maya, Revit, Rhino, and more, V-Ray is widely used for high-quality renderings in commercial applications such as architecture and product design.
V-Ray Benchmark is a free standalone application which allows users to evaluate hardware without having to install a full suite of software or provide a software license. For AMD GPUs, V-Ray Benchmark uses an OpenCL renderer, while for NVIDIA GPUs a CUDA-enabled renderer is used.
Interestingly enough, even with the newest drivers for the UHD 630 graphics on the i5-8400, the V-Ray Benchmark application couldn't find an OpenCL render device on our Intel test bed. Due to this limitation, we decided to compare the rendering time of the Ryzen 5 2400G to an entry-level discrete GPU, the AMD RX 550.
CPU and GPU rendering times cannot be compared with V-Ray due to the use different complexity scenes between the two tests.
As expected, the RX 550 provides a significant speed advantage over the Ryzen 5 2400G at about 3.2x. However, I am impressed with how well the integrated graphics on the 2400G were able to handle an intensive task like raytraced rendering especially comparing the price of $130 for just the RX 550 to $170 for the APU. Much of this performance gap is likely attributed to the slower memory sub-system on the Ryzen hardware (DDR4 vs GDDR5).
It's clear that AMD has an advantage over Intel Coffee Lake parts when it comes to more than just gaming on the integrated GPU. Any application that supports OpenCL acceleration should see additional performance from the more powerful Vega GPU. Real-world use cases are difficult to benchmark, but OpenCL acceleration is utilized quietly behind the scenes in lots of popular applications like Microsoft Office and Adobe's Creative Cloud programs.
Even for users looking to do more entry-level 3D rendering work, the Ryzen 5 2400G provides a great hidden value over the Intel Coffee Lake-powered CPUs.
Users will get the advantages of a quad-core processor with 8 threads for working inside the modeling applications, as well as the advantages of a capable GPU to render. Utilizing GPU rendering means that the system will remain responsive for light productivity tasks like web browsing, so you don't have to potentially lose the use of your computer while waiting for your rendering to complete.
Review Terms and Disclosure All Information as of the Date of Publication |
|
---|---|
How product was obtained: | The product is on loan from AMD for the purpose of this review. |
What happens to product after review: | The product remains the property of AMD but is on extended loan for future testing and product comparisons. |
Company involvement: | AMD had no control over the content of the review and was not consulted prior to publication. |
PC Perspective Compensation: | Neither PC Perspective nor any of its staff were paid or compensated in any way by AMD for this review. |
Advertising Disclosure: | AMD has purchased advertising at PC Perspective during the past twelve months. |
Affiliate links: | This article contains affiliate links to online retailers. PC Perspective may receive compensation for purchases through those links. |
Consulting Disclosure: | AMD is a current client of Shrout Research for products or services related to this review. |
Now we need to see the iGPU
Now we need to see the iGPU at 1600mhz
Blender 3D Cycles rendering
Blender 3D Cycles rendering done on the GPU, and Not on the CPU, is done via OpenCL for AMD and Nvidia uses the CUDA code path for Cycles rendering on its hardware in Blender 3D.
So for Cycles rendering done much faster on the GPU Blender 3D had mostly been a Nvidia/CUDA only option but a few years back AMD helped the Blender Foundation get Cycles working on AMD’s GCN, 2nd generation or later, GPUs! So AMD can offer better Cycles support now on these Vega based Graphics/Previous Generations of GCN(Excluding GCN 1.0) via OpenCL.
There is also AMD Pro Render plug-in for Blender 3D and I wish that Scott could be loaned as system to try out the R5 2500G for Cycles rendering(GPU only) and also that AMD ProRender plug-in(1). The ProRender users guide on page 9 lists the GPUs that can make use of the ProRender plug-in and that includes most of AMD GPUs including the RX and R9 series and other consumer variants in addition to the pro variants for windows 7, 8.1 and 10(Which is the only MS OS that supports the Radeon SSG variant). So that should also be looked at also under the ProRender plug-in manual(2).
So maybe Scott can be loaned some samples(2400G) and see if all the AMD support is up to speed for using on Raven Ridge Desktop, or even mobile, “APUs”. There is also Lnux support also to consider and AMD is really uped their sopport for all things Open Source these past few years.
(1)
“Radeon ProRender for Blender”
https://pro.radeon.com/en/software/prorender/blender/
(2) “Radeon ProRender plug-in for Blender User Guide v2.2”
https://www2.ati.com/relnotes/radeon-prorender-for-blender.pdf
I did do some Blender testing
I did do some Blender testing with Cycles on this GPU and it ran into some compatibility issues. Looks like there might have to be either a Blender or an AMD driver update for this to work properly.
I'll give the Radeon ProRenderer a look!
It looks like the usual
It looks like the usual driver/OS teething pains that every new CPU/GPU product goes through and I’m really wanting to begin to price out a system build for low cost Blender Rendering Mini-Desktop build on AMD’s Desktop Zen/Vega(“APU”) integrated graphics.
Looking over at Phoronix on the Linux side of things there appears to be more updating in store for the Linux Kernel for AMD’s Raven Reidge “APUs” also before the hardware has full support rolled out. I’m really not going to be using Ryzen 5 2500G for gaming that Much so if it gets to the point where integrated Vega’s Blender Cycles can a least support for GPU Cycles rendering via OpenCL Under windows or Linux then that would be good enough for my Blender 3D uasge.
Blender 3D is rare in that it makes use of OpenGL for its UI/Editor UI interface rendering and there is even a Blender 3D camera view port quick render/snapshot that’s done through OpenGL also. But to be able to make use of Blender 3D’s extensive Node(Materals nodes, compositor nodes, other nodes) capabilities that either has to be done on the CPU render path on Blender Render(CPU render, Slow) or Cycles on the CPU(Slow to render also). But Cycles is much faster done via GPU Cycles Rendering Path via OpenCL(AMD)/CUDA(Nvidia) render code paths for the rendering done on the GPU. And even Blender 3D’s Cemera View Port OpenGL rendered images take about one second compared to hundreds of seconds if I use the CPU to render.
It also looks like for AMD’s ProRender that under windows 10 only that Other makers’ GPUs may also work using ProRender and the Radeon ProRender Manual page 9 table lists Non-AMD GPU also with a checkmark under windows 10. Ubuntu® 16.04.3 is also supported for ProRender.
I think that AMD is probably working hard to get Raven Ridge support up to speed for gaming and Graphics software/other software usage with OpenCL GPU acceleration code path. So I hope that someone will keep testing the Raven Ridge APUs over the next few months more intensely like what was done for the Initial Ryzen 7, 5, 3 SKUs when they first became available.
The chips are ‘brand
The chips are ‘brand new’?
A.M.D. takes a little time to get stuff working best..
But they keep it working longer.
Worth waiting see what they come up with.
Just for reference with an
Just for reference with an Intel GPU where at last some effort has been made to get decent performance, namely Iris Graphics 540 (Surface Pro 4 in this particular case).
SiSoft Sandra 2017 – GPGPU Compute benchmark
Not too shabby considering 15 W TDP and a Q3 2015 launch.
finally a review telling
finally a review telling readers, that a benchmark company has not implemented ryzen sdk in it. a lot of game benchmarks dont have a ryzen patch, which is why the ryzen 7 8 CORE in SOME cases, still looses to an intel 4 CORE.
“Utilizing GPU rendering
“Utilizing GPU rendering means that the system will remain responsive for light productivity tasks like web browsing, so you don’t have to potentially lose the use of your computer while waiting for your rendering to complete.”
Since Chrome, Edge and Firefox (Opera as well I do believe) all use hardware acceleration for content rendering, wouldn’t that slow down your blender rendering??
Well edge sure beatschrome at
Well edge sure beatschrome at rendering for me.
I have a very weak amd 2 core laptop w/ HD 8240 gpu w/ 512 MB.
Since I started running edge concurrently (I cant face all the site password BS again) solely for video like youtube, i havnt looked back.
my cpu used to be constantly maxed out & gpu barely used, as per task manager.
Now, cpu maxouts are rare and brief, and dont cause any stutter, even at formerly impossible 1080p. It is clear from increased gpu usage, that it is being put to good use by MS Edge.
Very nice job, Ken!
I
Very nice job, Ken!
I especially like the “Review Terms and Disclosure”
something we just don’t see anywhere else on the Internet.
You guys at pcper.com are THE BEST, second to none!
/s/ Paul
You can download HW
You can download HW accelerated video transcoding software for Nvidia, Intel, AMD from here:
https://rigaya34589.blog.fc2.com/
For NVidia:
NVEnc v3.30
https://rigaya34589.blog.fc2.com/blog-entry-1006.html
For Intel:
QSVEnc v2.75
https://rigaya34589.blog.fc2.com/blog-entry-998.html
For AMD:
VCEEnc v3.06
https://rigaya34589.blog.fc2.com/blog-entry-901.html
As these are all CLI tools, you can find GUI for all these inside StaxRip v1.7.0.6 here:
https://github.com/stax76/staxrip/blob/master/changelog.md
Hope it helps.
use ffmpeg, the latest
use ffmpeg, the latest version has support for amf (amd hardware encoder), nvenc (nvidia hardware encoder) and qvs (intel hardware encoder)
also AMD APP SDK has been discontinued, it has been replaced by AMF
https://gpuopen.com/gaming-product/advanced-media-framework/
if am not mistaken StaxRip has support for all three
https://github.com/stax76/staxrip
SiSoft doesn’t look like it
SiSoft doesn’t look like it actually uses half precision on AMD, as that score should have basically doubled, not dropped… does it do this with dvega as well…?
The GPU encoding seems like it may be problematic too, as with dvega at least it can handle 4k60 at 100Mbps, assuming the encoder’s still very similar between ivega and dvega it should be more or less destroying intel there…
For comparison’s sake, specs
For comparison’s sake, specs from wiki:
UHD 630 Graphics in i5 8400:
192:24:3 @ 1050 MHz (403 GFLOPS)
Vega 11 in 2400G:
704:44:16 @ 1250 MHz (1760 GLOPS)
On paper UHD630 has around 23% of the performance of Vega 11.
Hi,
how long it take,
Hi,
how long it take, approximately, to repay a Rizen 2xxxG/MB/Ram combo minig cryptocurrency with it? Is It feasible?
Status update on my 2200g
Status update on my 2200g build… I should of waited for oem build, i think. I sort of knew in advance that I would have to get a loaner chip. Mobo came in and i put it in hoping it would work, nope. Emailed and requested a loaner chip from AMD, they sent a return email requesting a picture of my first born, copy of my receipts, and a note from mommy (mobo manufacturer) to tell why i need their loaner chip. Yikes, this in turning into quite the process for a cheap 100 apu. I understand though, AMD wants to keep the cost down and not get scammed by anyone. I suggest better comms with the motherboard manufacturers next time, perhaps drop ships from the maker directly if you order a specific need thing like this. Oh well…
Yet another update on the
Yet another update on the 2200g. What a pain this is, dont be a early adopter. Got the loaner, plug it in, no display. Could be ram incompatibility issues….? have to match ram to mobo and the CPU? This is getting ridiculous.
Hi whats the best gpu to
Hi whats the best gpu to partner with the Ryzen 2400G ?
Hi whats the best graphic
Hi whats the best graphic cards to partner with the Ryzen 5 2400G