Review Index:
Feedback

AMD Ryzen 5 2500U Mobile APU Performance with Raven Ridge

Author: Ken Addison
Subject: Mobile
Manufacturer: AMD

CPU and Gaming Performance

For our comparison points in this review, we are pitting the Ryzen 5 2500U-equipped HP Envy x360 against four other notebooks. First, we have the previous generation HP Envy x360 with the AMD FX-9800P APU. In addition, we have a series of three Acer Swift 3 machines: one with the last generation dual-core Intel Core i5-7200U, one with a Core i5-8250U quad-core, and another i5-8250U notebook, this time paired with an NVIDIA MX150 discrete GPU.

First, let’s take a look at Cinebench R15—a popular benchmark for CPU rendering performance.

View Full Size

View Full Size

While you aren’t likely to be rendering 3D scenes on a notebook like this, Cinebench gives us a good look at single vs. multi-threaded performance for a given processor. In Cinebench, we see an interesting split in results.

In single-threaded performance, the Ryzen 5 2500U falls short to the Kaby Lake Refresh i5-8250U processors by about 8%. However, in multi-threaded performance, we see the AMD 2500U with a 6% edge over the Intel parts.

View Full Size

View Full Size

Media encoding with X264 shows another mixed bag of results for AMD. While the Ryzen 5 2500U is slower than the i5-8250U in the first pass of encoding, the AMD part pulls slightly ahead in Pass 2. Keeping in mind that Pass 2 of encoding shows more of a performance benefit of multi-threading than pass 1 does, the R5 CPU seems to be slightly better at multi-threaded encoding tasks than the 8250U.

Geekbench 4 is a benchmark aimed at providing the overall performance of a CPU based on models of different real-world tasks.

View Full Size

View Full Size

Like Cinebench, Geekbench results are also separated into single and multi-threaded tests. Unlike Cinebench however, the Ryzen 5 2500U falls behind the Intel i5-8250U in single-threaded tests by 5% and in multi-threaded tests by over 25%.

With 3D rendering in Blender being one of the first workloads AMD showed off running on the Zen architecture, we expect to see the R5 2500U at its best here.

View Full Size

As predicted, the AMD R5 2500U-powered notebook finished first at rendering the BMW benchmark in Blender, by a margin of 3.5%.

Gaming

Mainstream gaming performance is one of the big targets for the AMD Ryzen 5 2500U processor. With integrated Vega 8 graphics, AMD has positioned this combined CPU and GPU part as a competitor for notebook solutions including Intel processors with low-end NVIDIA GPUs such as the MX150. To test this, we wanted to run our notebooks through a variety of titles.

  • Dirt Rally
    • 1600x900
    • Medium Settings, 2X AA
  • Rocket League
    • 1920x1080
    • High Settings
  • Counter-Strike: Global Offensive
    • 1920x1080
    • High Settings, 8X AA
  • Doom (2016)
    • 1920x1080
    • Vulkan API
    • Low Settings, FXAA

With a mix of low-end and higher-end titles, as well as a title that was demanding at its time of release, but has been out for several years (Dirt Rally), we hope to see a clear performance picture. We also decided to include Doom as a test of what advantages the Vulkan API gives on lower-end notebooks.

View Full Size

The Vega graphics in the R5 2500U processor completely dominate any integrated solution offered on Intel’s current generation of CPUs. We see a more than 2x gain in performance across our tested titles.

However, when compared to the MX150, the Vega 8 graphics seem to be no competition. The low-end NVIDIA discrete GPU offers as much as 50% better performance than the AMD integrated option.

The new AMD option also had some software compatibility issues (more on this later), which presented itself in the HP notebook consistently crashing when we tried to launch Doom, which is a flagship title for the Vulkan API.

Generational improvements

Even though the previous generation FX-9800P has been in all of our previous performance charts, we didn’t talk a whole lot about it in the context of competition between the current generation AMD and Intel parts.

While the FX-9800P was nowhere near competitive in any of our tests, it’s still a worthy data point to look at for context of the massive generational improvement that AMD has seen.

Comparing this new Ryzen and Vega-based APU to the last generation Excavator-core based APU shows massive improvements for AMD of around 3x. It’s impressive to see such a big performance jump in a relatively short amount of time, taking into consideration that the FX-9800P launched only around a year before the R5 2500U.


February 5, 2018 | 02:44 PM - Posted by bluezone (not verified)

Nice review. it would of been nice to see a frames per watt comparison in in the gaming portion. More frames usually consume more watts. NVidia discrete would still pull ahead though with that scenaro.

February 5, 2018 | 03:25 PM - Posted by remc86007

Thanks for the review and the disclaimer on the first page.

I would love some testing regarding how the CPU boosts clocks and what the thermal limitations of the chip are. Can two cores be at 100% load without power limit throttling if it's set to 15 watt tdp? Will it temporarily boost over the configured tdp for a short period of time like some of the intel parts do?

February 5, 2018 | 03:34 PM - Posted by AMDominance (not verified)

Nice to see the new disclosure :)

Any chance we could see benches that leverage this chip's inherent strengths in frequently used applications such as Adobe Photoshop? Perhaps balance it out with Lightroom as well if chipzilla frowns upon the results?

While this is a platform agnostic review, glad AMD has done a better job of getting appropriate platforms for their products out first - Carrizo was really gimped outside of the Lenovo y700 (and even there in some respects). That said, the 6bit panel on this model is, well, a 6 bit panel :/

I really REALLY hope Dell throws us a bone with an XPS 13 rocking this SoC and dual channel.

February 5, 2018 | 04:38 PM - Posted by CyclesRenderingTestedOnTheGPUPartOfAPUsPlease (not verified)

Yes good job PCPer with the transparency statements for reviews.

And PCPer is now Far Ahead of the entire online review industry for having a proper transparency statement included with its reviews. So those other review websites are now far behind PCPer in also including transparency statements with their reviews.

February 6, 2018 | 05:55 PM - Posted by John H (not verified)

Agreed, this is far beyond standard for the industry. Great job PCPer!

February 5, 2018 | 04:30 PM - Posted by CyclesRenderingTestedOnTheGPUPartOfAPUsPlease (not verified)

There needs to be some Integrated GPU ONLY Blender Cycles rendering done, Now more than Ever, so potential users can judge the Integrated Vega GPU's potential under Blender Cycles/OpenCL accelerated rendering on the GPU(Not the CPU).

I currently have an HP ProBook(Quad core Ivy Bridge i7) with a discrete mobile Radeon 7650M(Terascale Rebrand) GPU and my GPU can not do Blender Cycles Rendering on the Radeon GPU like can be not done on any Radeon 2nd generation GCN/later GCN based GPUs/integrated graphics.

So I know that the sepcific "Blender Render" on the CPU is a great benchmark for stress testing the CPU's cores but Blender's Cycles(GPU) rendering also works for GPU rendering via OpenCL on AMD's GCN 2nd generation/later GCN GPU micro-archs. So please do not forget to Test Blender's Cycles Rendering done on the APUs Integrated Vega GPU/graphics and not done on the CPU.

Blender Cycles Rendering(With the Settings Set to do the cycles rendering on the GPU via OpenCL) needs to be tested against Nvidia's Blender GPU accelerated cycles rendering(Via the CUDA code path) also.

I can not with my laprop's 7650M(Terascale Rebrand) do any Blender Cycles(GPU accelerated) OpenCL rendering so all Blender Cycles rindering on my probook has to be done on the CPU's cores. So I need to see more Cycles/GPU accelerated testing done on Raven Ridge APUs as a lot of laptops will come with only the Raven Ridge Mobile variants with only integrated graphics. And GPUs are a bit more performant when it comes to rendering than CPUs.

AMD's Driver Folks, along with the Blender foundation, put in a lot of work getting the drivers for Blender's Cycles rendering on the AMD's GCN(2nd generation/later GCN) GPUs working so Blender users could do Cycles rendering on the GPU accelerated via OpenCL. And most laptop reviewers are not Giving the Open Source Blender 3D users a fair amount of coverage for specifically testing Blender 3D on new laptops offerings.

There are plenty of Folks doing all sorts of Blender 3D workloads on laptops and AMD's APUs/Discrete GPUs since GCN 2nd generation have not gotten any attention as far as Blender GPU/Cycles rendering, mostly because of too much focus on CPU performance above all else.

AMD's intergated graphics have been ignored for too long with respect to Blender/Blender GPU/cycles rendering and that's not very just considering how performant AMD's graphics has become for the past few generations Polaris/later for OpenCL/Ray Tracing acceleration on the GPU.

So I'm Looking to upgrade to a Raven Ridge APU based laptop with a Mobile Raven Ridge variant or Better Yet I'm hoping that some Laptop OEM will offer desktop Raven Ridge APUs on a Laptop form factor, as that would be even better for some power laptop users.

From the Blender manual entry on GPU/Cycles rendering:

"GPU Rendering

Introduction

GPU rendering makes it possible to use your graphics card for rendering, instead of the CPU. This can speed up rendering, because modern GPUs are designed to do quite a lot of number crunching. On the other hand, they also have some limitations in rendering complex scenes, due to more limited memory, and issues with interactivity when using the same graphics card for display and rendering.

Cycles has two GPU rendering modes: CUDA, which is the preferred method for Nvidia graphics cards; and OpenCL, which supports rendering on AMD graphics cards.

Configuration

To enable GPU rendering, go into the User Preferences, and under the System tab, select the Compute Device(s) to use. Next, for each scene, you can configure to use CPU or GPU rendering in the Render properties.

CUDA

Nvidia CUDA is supported for GPU rendering with Nvidia graphics cards. We support graphics cards starting from GTX 4xx (computing capability 2.0).

Cycles requires recent Nvidia drivers to be installed, on all operating systems.

List of CUDA cards with shader model.

OpenCL

OpenCL is supported for GPU rendering with AMD graphics cards. (We only support graphics cards with GCN architecture 2.0 and above). To make sure your GPU is supported checkout this Wikipedia page.

Note

Cycles requires recent AMD drivers to be installed, on all operating systems." (1)

(1)

"GPU Rendering"

https://docs.blender.org/manual/en/dev/render/cycles/gpu_rendering.html

February 5, 2018 | 04:41 PM - Posted by CyclesRenderingTestedOnTheGPUPartOfAPUsPlease (not verified)

Edit: like can be not done
to: like can be done.

I hate word wrap with that text jumping between lines, but my proof reading sucks!

February 5, 2018 | 10:57 PM - Posted by SincereAnonymous (not verified)

I would love to see such a benchmark, since I don't ever recall seeing a Cycles GPU benchmark for APUs ever.

Although seeing as how this is still an integrated GPU and significantly slower than a MX150 even, I doubt anyone would seriously consider one of these for GPU rendering.

February 6, 2018 | 01:15 AM - Posted by NotForHeavyRenderingButSomeGraphics2Dand3DWorkOnALaptop (not verified)

No not for heavy duty GPU 3D animation rendering but it's still going to be faster done on the APU's integrated Vega GPU than on 4, Intel or AMD, CPU cores. So that's still going to be better than on my laptop's quad core Ivy Bridge i7 where some simple CPU based Blender rendering with a simple 3d model and the ray tracing sample rates set to the low-middle range settings can still take over an hour to render on 4 CPU cores/8 threads.

But if you are just doing a simple single scene/single frame image prohect then the APU's integrated GPU is going to be much faster. There are a lot of graphics arts projects that use 3D models and ray tracing for nice single image effects for single image designs. Even simple web animations can be created or 3D shadow effects for some 2D Graphics arts related projects where nice ray traced light/shadow effects can not be produced using 2d filters, or to create the effects using 2d methods/filters would take a couple of days but only 20 seconds on the mobile/integrated GPU if all the work was done by 2D methods using 2D graphics software.

Inkscape has a great feature where you can scan an image and have that image turned into an SVG with that generated SVG able to be done with from 2 to 256 scans, for 2-256 colors or grayscales, with each color or greyscale scanned producing a path/SVG(One path per color/greyscale up to 256 paths) and that can be imported into Blender and that 2D SVG turned into a 3D mesh. That's great for producing simple graphics projects like signs by then importing the SVG into Blender and turneing the 2d SVG into a 3d mesh. And setting up the lights and camera in Blender's 3d edit mode to render some nice drop shadow effects or glowing effects and creating nice graphics for web pages. So laptops are powerful enough for that using Intel's SOC or better AMD's APUs for those workloads.

Blender's grease pencil can be used to create 2D animations and all the backgrounds can be simple shallow 3D cutouts created in order for the animarion backgrounts to have depth and shodow/lighting effects and produce nice shadow casting to make the 2D animation appear to be more 3d like. Blender freestyle line rendering can be used to make any actual 3D objacts be rendered to look more like a 2D drawing. And all that simple sorts of projects can be done on a Laptop with APU graphics without needing a server room full of Rendering GPUs.

Blender's compositor and node editor can be accelerated via the GPU also and used to create many simple animation and single image graphics arts projects with related after effects on single image projects or simple 2D/3D animation projects. Even photographs can be brought into Blender's 3d editor and run through the commpositor or even assigned to a Mesh plane or shape for effects that can not be easily done on 2d graphics editing software. And all that can be done much better on the GPU than on the CPU. I have done some 2d work and just used Blender to render a photo that was assigned as a texture to a flat plane mesh and port that over to Blenders Node editor for some nice effects via the compositor nodes and used the GPU instead of the CPU and saved time.

AMD has those Radeon Rendering Plugins for Blender 3D also that can do some very nice Ray Tracing acceleration on the GPU. So for me any my needs the Vega APU(8 nCUs, 10 nCUs) graphics is probably a good deal better than my laptop's 7650M with its Terascale Rebrand Moble Radeon Graphics that can not be used for Blender cycles/GPU rendering at all.

What I'd really like is for a laptop to be offered with a desktop AMD Ryzen 5 2400G APU SKU inside and I'm also going to be looking at the Zotac ZBOX MA551 Mini-PC that offers AMD's Ryzen 5 2400G as an option for a mini desktop Blender rendering PC.

Laptops are fine for working up some single 3d mesh models and then creating materals and textures to assign to that single model that can be later incorporated into some larger game or animation project. So for Asset creation Laptops can be used as well as mini-PCs for single model creation and rendering to get the Model/its materals/textures created for inclusion in a larger projsct.

Laptops and any APU graphics are great for creating single mesh models and creating assets to be rendered on more powerful systems. And Laptops and Blender can be used for 2d work on simple 3D backgrounds for cutout backgrounds for 2D animation done via Blender's grease pencil and that animation plug-in that makes use of Blender's Greese pencel is great and even the cut-outs can be animated player piano style for moving animation backgrounds with the characters all hand drawn animations and the whole thing staged in Blender and edited via Blender's powerful node editor/compositior.

Blender's compositor and node editor(materals nodes, texture nodes, and other nodes) is like a little production studio with all the lights and cameras and other assets and plug-ins for doing all sorts of graphics workloads 2d and 3d or combinations of both and after effects croma-keyed in post production style. And a simple laptop with integrated graphics, or even low end disctete mobile graphics, is all you need for simple animation projects where some of the work is hand drawn and the backgrounds are combinations of hand drawn and actual 3D models, or simple cutouts with images mapped on them.

So Yes I want to see plenty of Blender testing done on AMD's Raven Ridge mobile/desktop variants and even some testing where some high polygon count scenes/models are brought in to Blender's editor just to see how many polygons can be worked on before Blender 3D's edit mode starts to bog down because the GPU's shader cores are starting to be taxed. And my current laptop's 7650M(terascale rebrand) starts to bog down if my mesh models/scenes get to around 1.5 to 2.5 million polygons total. Intel's integrated Graphics IvyBridge mobile SOC bogs down in Blender's 3D edit mode starting at about 300,000 polygons.

February 6, 2018 | 02:09 PM - Posted by msroadkill612

Yes, there is more to the new apuS than meets the eye.

Starting with a curiosity, that the "included free" gpu isnt just some long amortised gpu circuitry, but the almost unobtainable latest Vega.

Very fundamentally though, we have never seen such a promising interlinking of such diverse resources as in this apu.

For now, we see the cpu, gpu, memory controller linked - all bypassing the pci bus, interconnecting on the faster, smarter infinity fabric bus, over very small distances on a single die.

Add some APU integrated hmb2 cache, and it gets heady. HBM2 could serve like big L3 cache for the cpu it's so fast.

February 5, 2018 | 05:01 PM - Posted by CyclesRenderingTestedOnTheGPUPartOfAPUsPlease (not verified)

"There have been no updated drivers released for the HP Envy x360 since it’s launch in November, and likewise, owners of this notebook have been complaining to both HP and AMD about the situation. The community has figured out a (mostly) successful way to sideload the newer Radeon Software Adrenalin driver stack onto these notebooks, but that is an unsupported solution."

That's a long time problem with Laptop OEMs going all the way back to Intel's(In My Case) OEM(Toshiba) customizied HD Graphics drivers on my Toshiba C655 series laptop with a first generation core i3 CPU. And that Laptop is over 6 years old and never had a graphics driver update from the laptops's OEM. It't not Intel's fault as Intel can not Update non Generic OEM customized HD graphics drivers the laptop's OEM has to be the one to do that.

So file that Under #OEMCustomiziedLaptopGraphicsDriversGate because that's been happening for years and Laptop Reviews need to always call out the Laptop OEM's for customizing the GPU graphics drivers in the first place so the users are stuck between a rock and a hard place for graohics driver updates on OEM laptops.

February 6, 2018 | 07:46 AM - Posted by Anonymous12345 (not verified)

I’ve had plenty of amd notebooks and have never had an issue with pen driver locking a la intel drivers. If you look at the community workaround you can see that there is no ravenridge driver package at all in there.

February 6, 2018 | 10:19 AM - Posted by Bomsszoriz (not verified)

Well that's good to Know but on Intel's HD graphics based tat the craptular graphics drivers have to come from the Laptop's OEM(Toshiba) and Toshiba probably cheaped out on some features but so the OEM customizeid graphics drivers have to be updated from Toshiba, and that workaround(Side Loading generic Intel Graphics drivers) would not work as there was plenty of driver crashing when the Intel generic graphics drivers where side-loaded.

Hopefully for RR the drivers will come from AMD but laptop OEMs can never be tursted and look at all the AMD laptops that get the single channel to DDR memory treatment from Laptop OEMs.

I really hope that the System76/other Linux Laptop Folks get some offerings with AMD's revan Ridge APU options as that's where I will be on Linux come 2020 and windows 7's EOL. All my older laptops will be getting Linux and the newest one came with Windows 8 Pro(License) factory Down-Upgraded to windows 7 via the Pro version downgrade rights. So that Laptop will get windows 8, converted to 8.1 and some third party shell to turn all that TIFKAM OFF and make it look/act like windows 7 and that will last until 2023 befire 8.1 goes EOL.

"pen driver locking" ?

And the post was about laptops that all have Intel SOCs and some of the Laptop OEM's customizing the Intel HD graphics drivers on the Laptop OEM's products. Intel's Driver Update website would flat out refuse to update the OEM customizied Intel HD graphics drivers and any side loading of the generic Intel Drivers crashed too much. Please re-read the post that you replied to, re-read several times!

February 5, 2018 | 09:49 PM - Posted by Tiddums (not verified)

I'm interested to see how this stacks up for games that are heavier on CPU than they are on GPU. Stuff like Hearts of Iron 4 or Crusader Kings II.

February 5, 2018 | 09:55 PM - Posted by AcidSnow

It's amazing how I thought my HD 4000 iGPU was "OK" on this laptop I'm still using today. ...Sure I bought this piece of trash as a budget gaming rig, and it can play DotA and CS:GO with mostly decent frames (i5 3337U), but come on Intel - how are you guys so damn trash at integrated graphics? Not like I care, I like AMD more, and always have. But wow, how sad Intel must be.

February 6, 2018 | 12:22 AM - Posted by Anonymous7 (not verified)

cant understand the doom 0 , if this video clearly shows vulkan improvements in the last drivers

https://www.youtube.com/watch?v=-B6FlUA9PKI&t=19s

February 9, 2018 | 09:56 AM - Posted by Kareha

If you'd read the review then you would have noticed that Doom kept crashing when they were launching it, hence the 0.

February 6, 2018 | 04:36 AM - Posted by dragosmp (not verified)

Thanks Ken for another informative (5-in-1) laptop review. Since the TR review was published a while back it seemed this laptop told a muddy story - way faster than previous AMD, almost as fast CPU as current Intel with slightly worse battery life, or better battery life and somewhat worse performance vs Intel+Mx150. Probably about equal performance to Intel+MX130, but still lower power.

I want a light gaming laptop, but I don't know. I feel the Intel + MX150 is better in every way:
*better battery life for web/media when unplugged
*better CPU and GPU performance for when plugged
*same price

So why would I get the HP? Screen, keyboard, trackpad, speakers, they're all good(meh+) at this price point. Not a thing to sway a buyer towards the HP. Stepping back though, just the fast we think about AMD vs Intel in a 10mm-thick laptop and don't burst laughing, that's major evolution. The 9800P didn't feel slow at all, but this must be in a class above.

Can't wait for the Swift 3 w/2500U, I can only hope HP dropped the ball.

February 6, 2018 | 05:48 AM - Posted by JohnGR

About power consumption. No one expects AMD to get at Intel's levels with their first Ryzen models. That would imply that Intel engineers are second grade engineers at best, considering the years, money spend and probably the fact that Intel employs much more people in CPU design.
On the other hand there where negative comments about HP's laptop, and when compared the Ryzen x360 with the Intel x360, both laptops seemed to be bad at that power consumption category.
It seems that HP is not building only bad printers(I have one, I know).

February 6, 2018 | 08:55 AM - Posted by Spunjji

This right here - before drawing broad conclusions about Intel vs. AMD we need a way to eliminate the vendor variability. HP's Envy x360s seem to be extremely poorly optimized regardless of silicon vendor!

February 6, 2018 | 10:33 AM - Posted by Bomsszoriz (not verified)

AMD's Integrated Graphics is twice as powerful as Intel's dogfood integrated graphics so is not hard to see that AMD's Integrated vega graphics is going to use more power look at the shader coumts of the Vega integrated graphics. Most folks want the Raven Ridge more for the graphics than the CPU cores, and Zen is very close to Intel in CPU IPC for single core and better for all cores and That Vega graphics on an APU is what many want and who really cares about battery life, get a tablet if you want battery life over performance anyways.

AMD will be relesaing RR SKUs with lower nCU counts for the folks that want battery life above all else, but these RR SKUs currently have featers that Intel can not match graphics wise without a little help from Nvidia, or even AMD.

I want ASUS to field a Laptop with a desktop Raven Ridge APU and that APU SKU with Vega 11(nCUs) at 65 watts for a laptop as a desktop replacment usage. And that Asus with the desktop Ryzen 7 looks nice also. Really The desktop Raven Ridge SKU with the Vega 11(nCUs) and a Vega discrete Mobile GPU SKU sounds like a winner for graphics workloads on Blender 3d and Blender will happily make use of whatever GPUs are installed on a system for rendering on both the integrated and discrete mobile GPUs.

February 6, 2018 | 11:28 AM - Posted by Syed Wajih Hassan (not verified)

Firstly, commendations on the transparent disclosure. The industry needs to follow and improve upon. I understand the publishers have to operate a business, but such a practice demonstrates sincere intent towards the public/consumer. Very nice rebellion in todays' age of digital spies.

On the content, although higher power consumption, but very surprised to see the Ryzen parts matching CPU performance of 7th Gen i5

February 6, 2018 | 12:41 PM - Posted by willmore

Near the end of the page "Power Consumption and Software Concerns", you have "Core i5-2850U-based".

February 6, 2018 | 12:44 PM - Posted by willmore

Could you reconsider the way you run X.264? The two pass method has been depricated for years. The devs recommend (and most UIs for it) that you use the constant quality one pass method.

Good review, though. Nice to see a pretty balanced review. I'm typing this on an HP and I will add my voice to those saying that HP and power efficiency shouldn't be used in the same sentence.

February 6, 2018 | 01:49 PM - Posted by AMD Fan (not verified)

Thanks for a great review. I like the inclusion of a power consumption section as well.

I recently bought this laptop (envy x360) and I am very happy with it. The price point is something the article could be a bit more clear on. Bestbuy has this laptop for around $650. I paid another $160 for a 256GB samsung NVME M.2 drive and paid $40 for geeksquad to put it in. So, the total is around $850 without the sucky HDD.

Also, I dual boot both windows and arch linux. Using the laptop to compile the kernel is a pleasure. It never gets too hot to be on my lap and the fan noise (when it is turned on) is very low. Battery will drain faster for kernel compile but not crazy fast. I can still get a few hours out of it even all cores are 100%.

Using a bamboo ink pen with it in the tablet mode is also very nice if you don't mind the weight too much.

February 8, 2018 | 08:27 AM - Posted by msroadkill612

Thanks for a great review. :)

February 6, 2018 | 03:00 PM - Posted by Anonymous-911 (not verified)

How about adding a battery test based on Web Browsing/watching a movie to the mix. Also external GPU test?

February 8, 2018 | 03:11 AM - Posted by Abhola (not verified)

Is everyone taking Intel's meltdown performance hit into consideration? The landscape has changed considerably with Intel's performance advantage vanishing overnight.

February 8, 2018 | 08:38 AM - Posted by msroadkill612

Yes, it is all a bit strange. Oughtn't former benchmarks be revised to measure post patch performance for Intel CPUs?

AMDs too if you like, but intels aggressive use of prefetch contributes much to their much touted IPC edge. The post patch IPC difference would be less, imo

February 8, 2018 | 08:39 AM - Posted by msroadkill612

Yes, it is all a bit strange. Oughtn't former benchmarks be revised to measure post patch performance for Intel CPUs?

AMDs too if you like, but intels aggressive use of prefetch contributes much to their much touted IPC edge. The post patch IPC difference would be less, imo

February 9, 2018 | 04:36 PM - Posted by AMD notebook drivers - What a joke (not verified)

AMD will never get its act together re: notebook drivers.

This piece could've been written in 2005.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.