Ryzen APUs with Vega 8 and Vega 10

Shintai

Shintai

Supreme [H]ardness
Joined
Jul 1, 2016
Messages
5,678
2500U and 2700U have been spotted. Seems the official names for the IGP is now Vega 8 and Vega 10.
https://videocardz.com/newz/amd-radeon-vega-8-and-vega-10-mobile-spotted

AMD-Ryzen-5-2500U-with-Radeon-Vega-8-Graphics.png

AMD-Ryzen-7-2700U-with-Radeon-10-Vega.png


Ryzen 7 2700U compared to i5 7260U:
https://gfxbench.com/compare.jsp?benchmark=gfx40&did1=53156431&os1=Windows&api1=gl&hwtype1=iGPU&hwname1=AMD+Radeon(TM)+Vega+10+Mobile+Graphics&D2=Intel(R)+Core(TM)+i5-7260U+CPU+with+Iris(TM)+Plus+Graphics+640

upload_2017-8-21_12-17-47.png

apu.png
 
Must be those same codename Mandolin and Tambourine chips previously reported with naming scheme provided.
 
It is also confirmed that there is no HBM, that the L3 cache per CCX is one half that on RyZen (4MB vs 8MB), and that top APU model is 11CU.
 
It looks like Ryzen mobile does not have HBM2 on die and the chip is monlithic. I assume these will be fabbed in September at Glofo according to the Mosis schedule. According to AMD roadmap these will be BGA or LGA for mobile/embedded. Next year we will see AM4 and there is a possibility that HBM2 could be added to an MCM for the larger socket. Don't know yet.
 
gigaxtreme1 said:
It looks like Ryzen mobile does not have HBM2 on die and the chip is monlithic. I assume these will be fabbed in September at Glofo according to the Mosis schedule. According to AMD roadmap these will be BGA or LGA for mobile/embedded. Next year we will see AM4 and there is a possibility that HBM2 could be added to an MCM for the larger socket. Don't know yet.
Click to expand...

Neither mobile nor desktop RyZen APU have HBM.
 
Gideon said:
Your the only one that mentioned it on here. That seems to be the end goal for HBM, just not there yet.
Click to expand...

Yet you directly associate it with APUs? ;)

HBM on consumer graphics have been the biggest disaster ever so far. So how long will we have to wait on the promised land?
 
Shintai said:
Yet you directly associate it with APUs? ;)

HBM on consumer graphics have been the biggest disaster ever so far. So how long will we have to wait on the promised land?
Click to expand...
http://wccftech.com/xbox-one-may-be-getting-a-new-apu-based-on-amds-polaris-architecture/



Maybe I can find the AMD HBM paper talking about some wet dreams using HBM on everything, it was going to save the world too and put poor Nvidia out of business. All joking aside above source has to be considered as well.
 
Shintai said:
Yet you directly associate it with APUs? ;)

HBM on consumer graphics have been the biggest disaster ever so far. So how long will we have to wait on the promised land?
Click to expand...

Get over yourself you thought coffelake would be here on the 21st and you were wrong. Now it's turned into a #wait for Coffelake. You mentioned the whole HBM being on a APU no one said anything about it till you did. To me and others it seems like a possible use of HBM and would fix the biggest issue with integrated graphics. If Wccftech is your source for everything you think.. then well that explains alot. As for HBM being a disaster on video cards are you going to keep that tune since Nvidia uses it on their current Volta chips???? HBM works just fine for what it does and has nothing to do with with Vega's issues other then being delayed. I look forward to more spin with no facts to back them up.
 
Gideon said:
You mentioned the whole HBM being on a APU no one said anything about it till you did. To me and others it seems like a possible use of HBM and would fix the biggest issue with integrated graphics. If Wccftech is your source for everything you think.. then well that explains alot.
Click to expand...

Did you even check the link? It reports a roadmap appeared on a paper co-authored by an AMD graphics engineer

AMD-Zen-APU-With-HBM.jpg


Therefore the real source is an AMD enginneer.
 
lol, things don't always work out as planned do they?

The thing is to avert any disasters by recognizing the plan is not going to work and move in a better direction. Nvidia indicated they planned on using HBM for Pascal - that plan became obvious with issues, they worked with memory manufacturers to get DDR5(x) and soon DDR6, improved their designs and put out a superior product using old but improved tech. AMD just kept going straight for the cliff with song and dance to the drummer boy beat it seems.

Now that being said, if a Vega 64 LC was available for $699, I would most likely buy it :woot:
 
juanrga said:
Did you even check the link? It reports a roadmap appeared on a paper co-authored by an AMD graphics engineer

AMD-Zen-APU-With-HBM.jpg


Therefore the real source is an AMD enginneer.
Click to expand...

If you think Wccftech is a good source for info that is your issue not mine. I will not further waste my time with you.
 
Gideon said:
If you think Wccftech is a good source for info that is your issue not mine. I will not further waste my time with you.
Click to expand...

WCCFTECH is pure crap, but they aren't the source, WCCFTECH is only echoing something they found in internet. The source is a paper co-authored by an leading AMD engineer. Let me know if I have to explain this to you a third time.
 
Last edited:
gigaxtreme1 said:
You don't know that about desktop part for 2018. The older roadmap you quoted disagrees with you.
Click to expand...

I have known for a while that the Raven Ridge APU for the AM4 platform comes in 2018 with up to 11CU, doesn't have HBM but only dual channel DDR4, and has half the L3 amount per CCX compared to Summit Ridge. The older roadmap is wrong.

I have been saying this about RR, since when fantasy sites B&C invented 'rumors' about 16CUs and HBM2.
 
juanrga said:
Did you even check the link? It reports a roadmap appeared on a paper co-authored by an AMD graphics engineer

AMD-Zen-APU-With-HBM.jpg


Therefore the real source is an AMD enginneer.
Click to expand...


just because it says that doesn't mean it actually is, the internet creates stuff...i mean people crate stuff to post and start mass hysteria.
 
GFX bench guess it is better than nothing and the results are interesting given it is up against a 304 dollar IRIS part and unlikely to be anywhere near the same cost.

Looking forward to actual iGPU tests soon, could be an interesting part for a media centre.
 
OrangeKhrush said:
just because it says that doesn't mean it actually is, the internet creates stuff...i mean people crate stuff to post and start mass hysteria.
Click to expand...

Mass hysteria in the Internet has zero relevance to a roadmap given by an top engineer working at AMD.
 
The 7260u is 2/4 thead, correct? RR will still be the most capable desktop cpu overall with a 4/8 config.
 
juanrga said:
Mass hysteria in the Internet has zero relevance to a roadmap given by an top engineer working at AMD.
Click to expand...

is top engineer at AMD given a name or did the writer try sensationalism by claiming it came from a "top engineer". If said writer had dialog with said engineer the name may as well be provided, but it wasn't. This is what we refer to in a Court room as hearsay inadmissible evidence.

My uncles, daughter, friend heard from a friend who is Lisa Su's cousin that AMD is constructing the Death Star.
 
Seeing as how Shintai would never share leaks that paint AMD in a positive light, a large bucket of salt is required.
 
Nightfire said:
The 7260u is 2/4 thead, correct? RR will still be the most capable desktop cpu overall with a 4/8 config.
Click to expand...

Why do you want to compare a ultra-low mobile Intel chip with a desktop AMD chip? And are you aware that 4c8t Intel mobile chips are used in actual laptops? The 7260u isn't the fastest i7 by a large margin; it is not even the fastest ultra-low model.
 
OrangeKhrush said:
is top engineer at AMD given a name or did the writer try sensationalism by claiming it came from a "top engineer". If said writer had dialog with said engineer the name may as well be provided, but it wasn't. This is what we refer to in a Court room as hearsay inadmissible evidence.

My uncles, daughter, friend heard from a friend who is Lisa Su's cousin that AMD is constructing the Death Star.
Click to expand...

Cannot read that the real source is a paper coauthored by Mike Mantor?

http://www.amd.com/en-us/press-releases/Pages/michael-mantor-2015nov30.aspx
 
Last edited:
KazeoHin said:
Seeing as how Shintai would never share leaks that paint AMD in a positive light, a large bucket of salt is required.
Click to expand...

OrangeKhrush said:
At this point that doesn't even bother me, fanboys gonna fanboy it's life
Click to expand...

Where are then those leaks that paint AMD in a positive light? Share them with us, please. Or are you suggesting that him would invent leaks to please both of you?
 
juanrga said:
Why do you want to compare a ultra-low mobile Intel chip with a desktop AMD chip? And are you aware that 4c8t Intel mobile chips are used in actual laptops? The 7260u isn't the fastest i7 by a large margin; it is not even the fastest ultra-low model.
Click to expand...

Ask Shintai??
 
SimianRob said:
9vOBodI.jpg


3Dmark result
Click to expand...

Interesting and yet no comments on it from the usual people.
And to the people calling them fanboys, no, at least one in here is a shareholder shilling to boost share price.
Many [H]r's call a spade a spade. The extremely bias types are typically shareholders with skin in the game working PR, or employed to work forums, or just desperate fanboys. I'll happily rag on Intel or AMD or Nvidia when they are fucking up or being cunts. Doesn't bother me.
And before you call me an anti-intel shill, I have a certain contract with them in an unrelated field..
 
juanrga said:
Where are then those leaks that paint AMD in a positive light? Share them with us, please. Or are you suggesting that him would invent leaks to please both of you?
Click to expand...

Well this bench is good light, the physics performance is solid while the graphics score out does Iris Pro per 3DMark, given that all are 15w mobile parts that is quite an impressive showing.
 
OrangeKhrush said:
Well this bench is good light, the physics performance is solid while the graphics score out does Iris Pro per 3DMark, given that all are 15w mobile parts that is quite an impressive showing.
Click to expand...

Raven Ridge goes up to 35W on mobile and 95W on desktop.
 
You must log in or register to reply here.
Back
Top