AMD - Zen chitchat

[CAT-THE-FIFTH]

Btw,looking at that motherboard it looks like the one Bagnaj97 used for a build so should work fine with a Ryzen 5.

[Mr_Jon]

Btw,looking at that motherboard it looks like the one Bagnaj97 used for a build so should work fine with a Ryzen 5.

Confirmed. I have the 1600 in there.

[scaryjim]

Btw,looking at that motherboard it looks like the one Bagnaj97 used for a build so should work fine with a Ryzen 5.

According to the CPU support list it'll boot with any Ryzen CPU but only the non-X Ryzen 5 and 3 are fully supported in the original BIOS; the 1300X/1500X/1600X all require a update for full support.

But given there's been several AGESA updates since launch I'd probably have updated the BIOS ASAP anyway.

[Mr_Jon]

According to the CPU support list it'll boot with any Ryzen CPU but only the non-X Ryzen 5 and 3 are fully supported in the original BIOS; the 1300X/1500X/1600X all require a update for full support.

But given there's been several AGESA updates since launch I'd probably have updated the BIOS ASAP anyway.

The one I got last month had version 3.0 on it, from July, which is the second most recent. The latest is from mid-September.

[CAT-THE-FIFTH]

It seems Dell is making a RR based laptop called the Dell Latitude 5495:

http://ranker.sisoftware.net/show_sy...3f586bb83&l=en

[pp05]

I went for an ASRock B350 board, mostly for the USB 3.0 Type C on the back panel. Sure it lacks 3.1 Gen 2, but even if my stepson gets some g2 devices they'll still work in Gen 1 ports, so...!

I've only bought MSI/Abit boards in the past. So looking at reviews on it and how people are finding it.

[Bagnaj97]

I've just done a build with Asus Prime B350-Plus, Ryzen 5 1600 and 16GB (2x8GB) Crucial Ballistix Elite 2666MHz RAM. Board seems nice, with solid heatsinks on both parts of the VRM, and the ram is stable at 2933MHz 1.225v without dropping the latencies.

It was using the stock cooler and I didn't have much time, so I didn't play with OCing too much - things seemed stable at 3.7GHz, 1.2v but Cinebench15 scores were substantially lower than stock, so something wasn't right.

Cinebench scores:
Stock CPU clocks, 2666 MHz ram (16,17,17,36 timings) = 1145
Stock CPU clocks, 2933 MHz ram (16,17,17,36 timings) = 1150
3.7GHz 1.2v, 2933 MHz ram (16,17,17,36 timings) = 1052 - I think this can be safely ignored as it's clearly not right.

[Mr_Jon]

I've just done a build with Asus Prime B350-Plus, Ryzen 5 1600 and 16GB (2x8GB) Crucial Ballistix Elite 2666MHz RAM. Board seems nice, with solid heatsinks on both parts of the VRM, and the ram is stable at 2933MHz 1.225v without dropping the latencies.

It was using the stock cooler and I didn't have much time, so I didn't play with OCing too much - things seemed stable at 3.7GHz, 1.2v but Cinebench15 scores were substantially lower than stock, so something wasn't right.

Cinebench scores:
Stock CPU clocks, 2666 MHz ram (16,17,17,36 timings) = 1145
Stock CPU clocks, 2933 MHz ram (16,17,17,36 timings) = 1150
3.7GHz 1.2v, 2933 MHz ram (16,17,17,36 timings) = 1052 - I think this can be safely ignored as it's clearly not right.

The stock cooler with the 1600 is good, the fan lets it down a bit though in terms of noise...all my other fans are 140, so whenever the 92mm starting to speed up, I noticed it.

I attempted to strap a 140mm fan to the heatsink, but it's tricky due to the placement of the socket mounting screws. I dug out an old coolermaster hyper tx3 evo (that was seemingly destined for an easy life in a "spare parts" box), finding that it attaches to the stock AM4 mounting setup. A 140mm fan was then strapped, using cable ties, to the little trooper. End result is it's the most efficient cooler I've ever had. It keeps the stock chip under 50c during a full prime95 heat-orientated stress test, and quietly so.

Heat readings are a bit squiffy though, as there's more than one sensor at play. Speedfan and hwinfo report that top temp as 39c, whereas Ryzen Master and Speccy say it's 50. I go with the latter two. But all-in-all, the 1600 seems very easy to keep cool so there should be ample thermal headroom. I'm determined to find a way to mount a big fan on the stock heatsink too, to see how it performs. The big copper core means it should be a lot better than the little tx3.

[scaryjim]

... I'm determined to find a way to mount a big fan on the stock heatsink too ...

You don't seem to be able to get direct 140mm -> 92mm fan mount adapters, but you can get 140mm -> 120mm and 120mm -> 92mm ones, so you could build up a resizing duct from a couple of those. That would have the benefit of channeling all the airflow through the heatsink as well...

[DanceswithUnix]

You don't seem to be able to get direct 140mm -> 92mm fan mount adapters, but you can get 140mm -> 120mm and 120mm -> 92mm ones, so you could build up a resizing duct from a couple of those. That would have the benefit of channeling all the airflow through the heatsink as well...

Should be reasonably easy to knock up an adaptor with a 3D printer if you know someone who has one.

[Mr_Jon]

I bought an adapter off ebay that's flat, but with the sheer size difference between 80mm and 140mm (the heatsink is 80mm mountings, but a 92mm fan) there's a lot of material blocking the airflow. If I can't get a cable tie configuration going, I'll fabricate a proper alu bracket. The problem with the adapters is that it becomes hard to mount both the heatsink onto the board, and the fan onto the heatsink due to screw placements.

[DanceswithUnix]

This is the sort of thing I was thinking of, though this one is a bit tall and on that old heatsink I would probably have removed the AMD 70mm fan mounting clip and printed something that clipped straight to the metal. Bit harder to sketch but probably a better end result.

https://www.thingiverse.com/thing:470474

[Mr_Jon]

I got up and running with one. It's very tricky due to the mountings becoming obscured by the fan, but this worked: 80mm fan grill screwed onto the heatsink, 140mm grill cable tied to that, and then 140mm screwed to the 140mm fan grill. Heatsink and fan then mounted by using the screwdriver through the fan to attach the spring-screws to the backplate.

Results were utterly terrible. Ran the same test (where the evo bodge-up peaked at 50c), and stopped it at 74c (still slowly climbing). Probably would be better with a shroud, as the air was likely being pushed to the sides of the heatsink - the grills and ties created a gap between the fan and heatsink, whereas with the evo the fan was flush with it. Used a different 140mm fan (due to the frame obscuring the screw holes), but I doubt that is responsible for such a huge difference in temps.

[Xlucine]

You'd probably be better off with a 120mm if you aren't using a shroud - there'll be less dead air behind the motor

[CAT-THE-FIFTH]

First impressions of the HP Envy X360 with the Ryzen 5 2500U:

http://apusilicon.com/hp-envy-x360-w...t-impressions/
https://www.reddit.com/r/Amd/comment...ions/?sort=new

Edit!!

He has updated the results since he was running them with windows updated!! WTF??

Someone else also put up some benchmarks:

https://www.reddit.com/r/Amd/comment...ottom-comments


Cinebench R15 scores:

https://imgur.com/LOCjX2S
https://i2.wp.com/apusilicon.com/wp-...hr15udated.png

Cinebench R11.5 scores:

https://i1.wp.com/apusilicon.com/wp-...nchupdated.png

These two websites list comparative scores for the Core i5 8250U:

https://www.ultrabookreview.com/1780...0u-benchmarks/
https://www.notebookcheck.net/Intel-....242172.0.html

3DMark score:

https://www.3dmark.com/compare/is/4265626/is/4265355

[CAT-THE-FIFTH]

He tested Resident Evil 5 at 1080P at max settings:

https://www.youtube.com/watch?v=0A0Y...ature=youtu.be



https://www.reddit.com/r/Amd/comment...sions/dpsvthq/