AMD - Zen chitchat

[scaryjim]

... I would be useful if you could give us some first impressions,especially regarding the IGP and GPU!!

After all the fuss I've made over it the least I could do is a quick user review

[CAT-THE-FIFTH]

After all the fuss I've made over it the least I could do is a quick user review

At least you get a better warranty than the grey importer!!

[CAT-THE-FIFTH]

Looks like SiSoftware Sandra results were leaked early for the Ryzen 5 2600 and Ryzen 7 2700X:

https://www.reddit.com/r/hardware/co...2700x_testing/

Also,LOL at the AMD marketing:

https://imgur.com/a/3gnMf

[DanceswithUnix]

Anyone played with a Gen10 microserver? Missed that these were AMD based, shame they use a soldered down part so you can't upgrade to a zen based apu.

https://www.serversdirect.co.uk/p/12...-x-non-hotplug

In my new job I need a small cheap server to store a few files and run a few VMs so that looks about right.

[CAT-THE-FIFTH]

Asus ROG X470 pictures:

https://www.techpowerup.com/242439/a...out-of-nowhere

[CAT-THE-FIFTH]

More Asus X470 details are leaked:

https://videocardz.com/75433/asus-x4...ve-been-leaked

The max supported RAM is 3466MHZ!

[CAT-THE-FIFTH]

FFS,looks like Dell really seems to not care about its Ryzen laptops:

https://www.extremetech.com/computin...l-counterparts

Also LOLWTFBBQ on the Ryzen 3 2200U based ones.

Edit!!



Hardware Unboxed shares how their Ryzen system fared over 12 months of usage.

It also appears some of the launch "memory problems" were more down to the early Asus ROG motherboards supplied for launch having faulty RAM slots!!

WTF??

[scaryjim]

... I would be useful if you could give us some first impressions ...

First impressions up at https://forums.hexus.net/reader-revi...eon-530-a.html

[watercooled]

More evidence of the CTS thing being obviously malicious in some way? https://www.anandtech.com/show/12556...tched-in-weeks

Apparently it's all fixable and in a short space of time (weeks, not months) i.e. likely could have been fixed before anyone else knew about it if they'd done the sane thing and allowed AMD a reasonable time to respond.

[CAT-THE-FIFTH]

The French magazine CanardPC Hardware will be featuring a full Ryzen 7 2700X review on Monday!
https://videocardz.com/75489/amd-ryz...view-on-monday

Heads up to those in France. On Monday everywhere in France (and on Saturday in Paris) a new issue of CanardPC Hardware magazine will be available. This magazine will feature two Ryzen 2000 reviews, the Ryzen 2200G/2400G and also Ryzen 7 2700X, which is launching on April 19th.

While this is all interesting, and we can’t wait to see the results, this is in fact, the first picture of retail Ryzen 7 2700X to be made public. No other details were revealed, let’s hope it was tested on X470 motherboard as well.

CPC:

We have found significant improvements on some points … but also an unpleasant surprise that we let you discover!

I hope its not tested in an A320 motherboard as the cover picture indicates!!

Edit!!

The magazine might be out tomorrow in Paris.

[CAT-THE-FIFTH]

Comment on the article:

https://translate.google.co.uk/trans...zen-7-2700x%2F
https://www.reddit.com/r/Amd/comment...n_2000_series/

The PC duck April-May 2018 has been delivered to subscribers since yesterday. And this number is a bit special, since it includes the test of Ryzen 7 2700X, Ryzen 5 2600X and Ryzen 5 2600.

We will not show you the graphs of the results, for a reason of respect. However, we will talk about the trend that emerges.

The Ryzen Pinnacle Ridge on Tun Turbo was better scaled than the Summit Ridge, which saw their frequency drop as soon as more than two hearts were used. The Pinnacle Ridge, including the Ryzen 7 2700X therefore, also have a "modified" uncore (to understand by that he repaired what was working badly), which leads to a better latency of caches and memory. In addition, it is now possible to clock the core and the uncore in a differentiated way. But what does it give?

In games the Ryzen 7 2700X is of course better than the 1800X, but still does not reach the level of Hexacore Intel (Core i7 8700K in mind). By cons in application, AMD has a sacred advantage. Consumption level, it's a bit cold shower since the latter takes off against the old generation (equivalent processors: 1600X vs 2600X, 1800X vs 2700X ...). The engraving "12 nm" (note the quotation marks) does not make miracles either: the ratio perf / conso does not evolve. It was pretty good on the first generation, as we showed in our test of Ryzen 7 1800X , Ryzen 7 1700X and Ryzen 7 1700 .


Their magazine will be available from Monday on the shelves. Go buy it, it's worth the detour and the eternal pink rabbit is always there!

It appears an A320 based motherboard was used!!

[watercooled]

Hmm, I know it's not a fantastic translation but this part caught my eye:

The small nuance to bring is the fact that CPC Hardware had to use an A320 motherboard, with an AGESA 1.0.0.1a. Which implies that the CPUs in question had their Precision Boost disabled, this technology being only on the 400 series cards. This does not change the performances much, 1 to 2% maximum maximum, and does not change the ratio perf / conso in half-tone.

Only that's not necessarily true. I'm not too sure of the specifics of how Precision Boost improves upon its predecessor but more granular and/or faster power state management can substantially change both performance (particularly in bursty workloads) and efficiency. If that were not the case, then neither AMD nor Intel would be investing as much into it as they are - overcoming power delivery challenges and getting fast clocks ASAP to complete tasks, then getting those parts gated ASAP after completion is exactly what they're there for.

[Xlucine]

Precision boost 2.0 = higher all-core speeds (supposedly):

https://hexus.net/tech/news/cpu/1160...ap-specs-leak/
Not a proper graph, so we can't quantify the claimed 1-2% max performance difference yet

I wonder if that A320 board has the VRMs to properly supply the XFR? It's going to be an interesting week once more reviews launch!

[watercooled]

I thought (perhaps incorrectly) that there was more to it than that, like with the Ryzen APUs vs 1st gen Ryzen? Things like increased granularity, faster response, etc - they don't include features like that for nothing; it's either power, performance, or both.

Having said that, even 100MHz is more than 1-2%!!!

[Xlucine]

Precision boost 2 is (part of) the more granular bit, per the mobile ryzen slide deck:

https://hexus.net/tech/news/cpu/1114...vega-graphics/

They call it 2.0 in the leaked slides, and just 2 in the real older slides, but it's probably the same thing. What's also interesting is that for the mobile slide deck, they were talking (in general terms) about PB2 not resulting in a higher all-core boost:


We might be looking too much into some conceptual graphs, of course

ETA: also, a 25 MHz boost is only 0.6% for a 4 GHz processor

[scaryjim]

PB2 should allow cores to run at higher clock speeds when they're not heavily loaded - so you might have a workload that uses lots of threads but only one or two are particularly intensive (I know a few game engines do that - they have a couple of intensive threads then many light load threads). PB2 should benefit those loads by intelligently boosting cores that aren't as heavily loaded.

Ryzen 2 'X' processors seem to get an XFR enhanced version of PB2 which will also assess the cooling capacity and attempt to maintain higher boost clocks than normal.

Both of those technologies are meant to work in any motherboard.

Using a 400-series motherboard with a Ryzen 2 'X' appears to allow a higher level of multi-core boost clocks (under this 'Precision Boost Overdrive' nomenclature), so a 2700X in an A320 motherboard should be slower in heavily threaded tasks than the same CPU/Cooler in an X470 board.

How much of all that is just marketing bumf, OTOH, remains to be seen...