AMD Ryzen 5 2400G and Ryzen 3 2200G APUs performance unveiled

[xblax]

Everybody was talking about games and someone said that the i5 would outperform the APUs, of course he was talking about games as well.
Also anything other then games that uses all cores will run much faster on a dGPU then on the additional cores of the APUs.
Here for example blender 3d rendering
GTX 1080: 11.24 s
https://forums.anandtech.com/thread...-amd-blender-benchmark.2507954/#post-38923283
AMD ThreadRipper 13.04 seconds
https://forums.anandtech.com/thread...-amd-blender-benchmark.2507832/#post-38920471
Running "productivity" apps on CPU has stopped being a good idea for years now,GPUs are just so much faster.

For 3d rendering you're right. But there are still many workstation tasks that don't benefit from GPUs. For compiling, datbases, virtual machines and even for high quality video encoding GPUs are useless. Most developer workstations would be fine with Threadripper + GT1030/RX550 and don't need a powerfull GPU at all.

For 2200G/2400G price/value comparison I can only say that I would prefer the 2400G for what I'm doing on my desktop any time. However, I'm still undecided between APU or CPU+dGPU build with Zen+. Not because I really need a dGPU but because I allready have one.

 

[Shivansps]

This guy has one and is running games and benchs, im watching right now, ill post the time of the interesting things.

He only managed to boot up at 2667mhz, so that will petty much kill the gaming.
oh hell again that bsod with the radeon driver, and haha he dindt know you have to force the IGP to always on to be able to change the shared memory on some bios.

Games starting at 4:30:00

WTH is wrong with AMD? they plan to do a worldwide launch of RR tomorrow and they still havent put up a working RR driver?

 

[Kenmitch]

That video sucked....Too bad I can't type what I really thought about it here.

Dissing AMD for not having drivers online for a product that hasn't been released yet seems silly at best.

The reviews will be interesting once online.

 

[Shivansps]

That video sucked....Too bad I can't type what I really thought about it here.

Dissing AMD for not having drivers online for a product that hasn't been released yet seems silly at best.

The reviews will be interesting once online.

For a product that releases in HOURS. That is like saying "whats the point of offering firmware updates to support it", they did release it. They know by now there are people with the APU and this makes them look bad. Press should have a working driver for days now, whats the point of hidding the driver? Specially because the OEMs are offering a driver that supports it and its completely broken.

The good news is that 17.40 is just completely broken for RR IGP, and so a new driver could provide a nice boost in the scores we were seeing for the last few days.

 

[Kenmitch]

For a product that releases in HOURS. That is like saying "whats the point of offering firmware updates to support it", they did release it. They know by now there are people with the APU and this makes them look bad. Press should have a working driver for days now, whats the point of hidding the driver? Specially because the OEMs are offering a driver that supports it and its completely broken.

It's not really my concern. If somebody that doesn't have the proper drivers and wants to post youtube videos that make themselves look bad why would I care.

I don't really see the point of complaining about the drivers before the product is released....Unless it's just to downplay, diss, complain, etc about AMD.

 

[yforu]

Dissing AMD for not having drivers online for a product that hasn't been released yet seems silly at best.

There is still no official driver from AMD for Raven Ridge two month after the release. So yes, he is right about that. Don't forget mobile APUs

 

[Kenmitch]

There is still no official driver from AMD for Raven Ridge two month after the release. So yes, he is right about that. Don't forget mobile

Your saying that the OEM's are selling laptops and not even giving you drivers for it? I'd be mad if that was the case....Somehow I think it's not the way it works.

 

[Shivansps]

It's not really my concern. If somebody that doesn't have the proper drivers and wants to post youtube videos that make themselves look bad why would I care.

I don't really see the point of complaining about the drivers before the product is released....Unless it's just to downplay, diss, complain, etc about AMD.

This should be a subject we should all agree on, unless you feel some need to defend every AMD move even when they make no sence at all.

No one said it should be final fully optimised drivers, but at this point there are people, like the one on the video with the APU and they start to give information based on a broken driver, while they sit on a driver that actually works waiting for the clock to run out. Whats the point? there is none.

What it puzzles me is that they had to give the go ahead for OEMs to release 17.40 driver dated from january for RR desktop, petty much like they do for ANY dgpu, they had to knew that was completely broken. Even more reason to release a driver sooner.

 

[yforu]

No support for Vega IGPs with Adrenalin Edition 18.2.1 (2/7/2018). There is just the old driver from HP (11/20/2017) for Ryzen Mobile. Official support -> AMD Driver Downloads.

 

[xblax]

This should be a subject we should all agree on, unless you feel some need to defend every AMD move even when they make no sence at all.

No one said it should be final fully optimised drivers, but at this point there are people, like the one on the video with the APU and they start to give information based on a broken driver, while they sit on a driver that actually works waiting for the clock to run out. Whats the point? there is none.

What it puzzles me is that they had to give the go ahead for OEMs to release 17.40 driver dated from january for RR desktop, petty much like they do for ANY dgpu, they had to knew that was completely broken. Even more reason to release a driver sooner.

It's definitley a bit weird that there are no Raven Ridge drivers even for mobile chips on their website at all. But if an updated driver surfaces tomorrow it's still going to be ok. I hope that testers didn't receive the same outdated 17.40 driver. All mainboard vendors seem to have 17.40.3301 online so far and it's distributed in combination with the chipset drivers.

 

[tvdang7]

can some one explain something to me. How are these processors any different from the ones that are coming april?

AMD has been doing a great job at keeping the average Joe confused.

we had APU's and FX cpus.
we got line up that consists of RX580 and a product line called vega.
Now we got Ryzen 2 apu and Ryzen 2(or+) coming? and The REAL Ryzen 2 is coming next year.

 

[coercitiv]

AMD has been doing a great job at keeping the average Joe confused.

Now we got Ryzen 2 apu and Ryzen 2(or+) coming? and The REAL Ryzen 2 is coming next year.

You got things a bit backwards: next Zen architecture is coming next year, Ryzen is the commercial brand just like Core is Intel's brand.

 

[rainy]

can some one explain something to me. How are these processors any different from the ones that are coming april?

AMD has been doing a great job at keeping the average Joe confused.

we had APU's and FX cpus.
we got line up that consists of RX580 and a product line called vega.
Now we got Ryzen 2 apu and Ryzen 2(or+) coming? and The REAL Ryzen 2 is coming next year.

Pinnacle Ridge (also Ryzen 2xxx series) it would be an improvement over Summit Ridge released last year and like predecessor it don't have IGP.
Next year we will see Zen 2 (Matisse aka Ryzen 3xxx series) with some serious architecture changes.

 

[The Stilt]

WTH is wrong with AMD? they plan to do a worldwide launch of RR tomorrow and they still havent put up a working RR driver?

So you recon the only way for the reviewers to get the drivers is to wait until AMD officially releases the package at their support site?
Please.

The guy who posted the video doesn't have the proper drivers because he isn't a reviewer nor supposed to have the chip at hand in the first place.
He got it early from an etailer by an accident.

 

[Shivansps]

So you recon the only way for the reviewers to get the drivers is to wait until AMD officially releases the package at their support site?
Please.

The guy who posted the video doesn't have the proper drivers because he isn't a reviewer nor supposed to have the chip at hand in the first place.
He got it early from an etailer by an accident.

I want to belive AMD provided more decent drivers to reviewers sooner than that. What that guy is using is what every OEM is right now providing since Feb 6th as "RR Support" what i do belive is some early beta that should have never seen the light of day, for AMD own sake.

 

[The Stilt]

I want to belive AMD provided more decent drivers to reviewers sooner than that. What that guy is using is what every OEM is right now providing since Feb 6th as "RR Support" what i do belive is some early beta that should have never seen the light of day, for AMD own sake.

Of course they did and the bioses as well.

 

[VirtualLarry]

Of course they did and the bioses as well.

I question that. ASRock still lists the newer UEFI for the AB350M Pro4 as 4.50, as the BIOS for the "Next-gen Ryzen APU", yet, that BIOS/UEFI is buggy as F. Can't even boot Windows 10 properly with my Ryzen 5 1600 with it.

So, from my perspective, we've got BIOSes out for RR, that DO NOT WORK RIGHT with existing 1000-series Ryzen CPUs, will we have the same situation with Sandy and Ivy that Asus Z68 boards had, that in order to run an IVB in those boards, you had to flash the "updated BIOS", but then you could NO LONGER run Sandy CPUs in them, without putting in an IVB and flashing back, and shutting down to swap and re-start?

This launch is a MESS!

 

[The Stilt]

I question that. ASRock still lists the newer UEFI for the AB350M Pro4 as 4.50, as the BIOS for the "Next-gen Ryzen APU", yet, that BIOS/UEFI is buggy as F. Can't even boot Windows 10 properly with my Ryzen 5 1600 with it.

So, from my perspective, we've got BIOSes out for RR, that DO NOT WORK RIGHT with existing 1000-series Ryzen CPUs, will we have the same situation with Sandy and Ivy that Asus Z68 boards had, that in order to run an IVB in those boards, you had to flash the "updated BIOS", but then you could NO LONGER run Sandy CPUs in them, without putting in an IVB and flashing back, and shutting down to swap and re-start?

This launch is a MESS!

How did you determine that the issue is related to code from AMD side?
There are several boards which work just fine with the same software stack (RavenPI or PinnaclePI) from AMD, on existing parts (e.g. ASUS Crosshair VI Hero, Gigabyte AB350N-WIFI, MSI B350I PRO AC).

EDIT: I had a look at the bios you mentioned (AB350M Pro4 4.50). The bios isn't even based on the most recent code path.

 

[3DVagabond]

I question that. ASRock still lists the newer UEFI for the AB350M Pro4 as 4.50, as the BIOS for the "Next-gen Ryzen APU", yet, that BIOS/UEFI is buggy as F. Can't even boot Windows 10 properly with my Ryzen 5 1600 with it.

So, from my perspective, we've got BIOSes out for RR, that DO NOT WORK RIGHT with existing 1000-series Ryzen CPUs, will we have the same situation with Sandy and Ivy that Asus Z68 boards had, that in order to run an IVB in those boards, you had to flash the "updated BIOS", but then you could NO LONGER run Sandy CPUs in them, without putting in an IVB and flashing back, and shutting down to swap and re-start?

This launch is a MESS!

Maybe I don't understand. This seems like a non issue. It's not like you are going to switch back and forth between CPU's and need support for both on your MB all of the time.

 

[alexruiz]

I question that. ASRock still lists the newer UEFI for the AB350M Pro4 as 4.50, as the BIOS for the "Next-gen Ryzen APU", yet, that BIOS/UEFI is buggy as F. Can't even boot Windows 10 properly with my Ryzen 5 1600 with it.

So, from my perspective, we've got BIOSes out for RR, that DO NOT WORK RIGHT with existing 1000-series Ryzen CPUs, will we have the same situation with Sandy and Ivy that Asus Z68 boards had, that in order to run an IVB in those boards, you had to flash the "updated BIOS", but then you could NO LONGER run Sandy CPUs in them, without putting in an IVB and flashing back, and shutting down to swap and re-start?

This launch is a MESS!

What exactly is what "doesn't work right"?

 

[Abwx]

There s no "true" AGESA supplied by AMD, what they give are micro code building blocks that are assembled by the Bios manufacturer (AMI FI) in function of the MBs manufacturer s needs, this way AMD can supply an updated memory management module (!) FI without the rest of all these specific bios being touched by this specific update, all the OEMs have to do is to use the same previous bios with just this little routine being updated...

 

[coercitiv]

https://www.asrock.com/mb/AMD/AB350M Pro4/index.asp#BIOS

Weird way to perform the update, having to flash a "bridge BIOS" first. Then again @VirtualLarry can't say he wasn't warned

We don't recommend users to update the BIOS if their system is already running normally

 

[Abwx]

Some AM4 MBs (the most expensive of course..) can be updated using USB and without a CPU being plugged...

Otherwise, well, there some cheap Bristol Ridge for the purpose...

 

[VirtualLarry]

I'm just saying, ASRock has a mini-site calling out support for "Next-Gen Ryzen APUs", and they specifically mention the AB350M Pro4, the 3.40 "Bridge BIOS", and the newest 4.50 UEFI as "the" BIOS for Raven Ridge, and it's garbage for 1000-series Ryzen CPUs.

I literally couldn't even boot Windows 10 using that BIOS, or it would blue-screen shortly after booting. It was THAT bad. And yes, I was careful to clear CMOS, both before and after flashing the BIOS. (Well, in UEFI, selecting "Reset to UEFI Defaults".)

From what I can see, that new AGESA is garbage, or ASRock's BIOS engineers are (which hasn't been my experience).

I bought a Gigabyte B350 board, recently, for my Raven Ridge adventures. I don't trust ASRock not to botch things.

 

[chrisjames61]

Are these apu's going on sale tomorrow in the USA? Anybody know?