Hacker News Reader

Comment by xedarius | Hacker News Reader

I recently bought a Ryzen. I was moving from a mac to a pc and I hadn't owned a pc in over 10 years. Thing is, 10 years ago my pc had 4 cores .. and at the time of choosing Ryzen it was the only consumer processor (with a reasonable price tag) that was offering more cores.

I'm very pleased, my Ryzen is fast and stable, I'm glad the company is reaping the rewards.

View on HN

Another Ryzen owner here. I came from mobile 2nd and 3rd gen i7s.

The price to performance is insanely good. I'm on an x370 desktop, with 2xrx550 and rx560. I'm using it as a virt/dev host. I have 16 docker instances running on the host os (Arch Linux).

Then 3 vms with PCI pass through. One for windows gaming/development. Until about three weeks back I was getting a solid 60fps @ 1080p on high across most games, but recently it's been down below 10fps. The next VM is a Linux developer desktop. Lastly, an emulator box connected to a projector away from my desk.

To switch between I just have dongles, 1xhdmi and 1xusb, behind my keyboard, and alternate the plugs. I did this because I couldn't find a good hdmi kvm that did 4k above 60hz.

This year has really been amazing for enthusiasts.

Why has the framerate dropped so much?

I have no idea, to be honest. I tried different driver versions in windows, host kernels, etc. Even booting back into native windows OS, no KVM in between. It was showing the same frame rate drops. My 3DMark was also just about halved as well.

Sounds like an awesome setup :) How much load does your cpu have running so many containers and vm's? I have heard of software solutions for io switching. So you could get rid of usb switching.

I run an Arch host with passthrough myself but it was really a long project setting everything up just right. Especially for gaming.

Really looking forward to buying a Ryzen or Threadripper to go further and virtualize my working enviroment etc.

External public link, I also just realized my monitoring isn't fully active but you can see historical trends.

https://animusnull.drunkensailor.org/grafana/dashboard/db/no...

The CPU is barely breaking a sweat. I wouldn't mind a thread ripper so all VMS can get a quad core. But right now it's plenty powerful.

The input was the biggest pain in the butt. Especially on windows... I've seen the evdev passthrough and hot swap but didn't have the best luck with it. While hacky. Assigning a USB controller to each VM has worked best.

I agree it's a long setup, and still requires some care. It's not straight up and running like a linux/windows desktop. I've had windows just mark the usb controller as inactive, and I then need to reboot the entire host. Something I'm still digging into. But there is much better documentation out there. Things like project looking glass are also impressive.

> This year has really been amazing for enthusiasts.

I mean, graphics card prices have been pretty horrific, but I agree on the CPU side.

RAM prices are through the roof as well. SSD seems to finally be stabilizing somewhat though.

Right? I can’t wait for Bitcoin to crash. Let us do something of real value than mining coins.

Awesome, what do you use to run the VMs ?

Only KVMs that I'm aware of that'll work for this are expensive and display port rather than HDMI (not a big deal, passive adapters exist).

https://store.level1techs.com/products/kvm-switch-single-mon...

https://store.level1techs.com/products/kvm-switch-dual-monit...

Yea the whole zen lineup seems to have been huge for the company. I switched from an older intel platform (i5 4690k) that i was using to host a bunch of virtual machines for development and play to a threadripper. the difference is night and day, and it's an affordable way for me to go to ECC ram since I've got the need for a lot of it with my use case. Very happy to see AMD making great strides again.

Oh, what's your use-case for ECC ram?

I'm not the parent commenter but it's probably finance or CAD modeling, environments where soft bit-errors (ones that silently and unpredictably cause data corruption rather than hard-crashes that are reproducible) can lead to nightmares.

Or could be using ZFS

"There's nothing special about ZFS that requires/encourages the use of ECC RAM more so than any other filesystem." - Matt Ahrens[0]

[0]: https://arstechnica.com/civis/viewtopic.php?f=2&t=1235679&p=...

Except there is:

https://forums.freenas.org/index.php?threads/ecc-vs-non-ecc-...

I didn't say it was required. It's not uncommon for people to want to use ZFS specifically because they highly value data integrity. If that's the case, then it probably makes sense to use ECC.

The good question would be what's use-case for non-ECC RAM. I wouldn't want unreliably hardware for any task, even as simple as gaming. It's a shame that ECC ram available only for enthusiasts and with unnecessary premium. ECC really should be baseline. Computers meant to be reliable.

Is there any windows / linux utility that shows the number of ECC errors corrected since the boot?

If not, AMD should write and promote one.

Very curious on how often it happened for normal home/offic usage.

I used to work for a silicon company who took a embedded network switch system with ECC logic to some nuclear lab for testing to verify/showcase the ECC functionalities.

Even if you never personally encounter an error, knowing that the 1 bit correct and 1+ bit likely detect exists will save you in terms of piece of mind and trouble-shooting other issues.

Plus, if you're scrubbing your storage the last thing you want is a memory error killing your data.

In Linux, yes, there is a service called mcelog and a utility from the edac-utils package called edac-util.

You will see correctable ECC errors on systems. How frequent honestly seems to depend on the workload and the system itself. My suspicion is that often they are caused by poor PCB layout and ECC saves you. I spent literally weeks (nights, weekends) chasing down an issue I thought was a software bug but turned out a board layout issue on an embedded system. If the system had ECC, the error would have either been corrected or we would have gotten the uncorrectable ECC error trap. Since then, every workstation/server/desktop I spec is ECC. I wish more laptops had it.

Thanks,

Try the edac-util on 5 of 30+ or so servers.

   "Intel(R) Xeon(R) CPU E5-2667 v3 @ 3.20GHz" with 128G RAM each. 

   edac-util: No errors to report.

System uptime ~30+ days. We recently have to move those servers. I wish I check this command before the move. The uptime should be 600 days + for some of the servers.

The problem with bit flips is that they accumulate in high uptime systems.

If you reboot your PC at least once every week it's not going to be a problem.

Early ECC days, you 'washed' memory to fix this. On a read, a single-bit ECC error is actually repaired by the hardware. To get the most benefit from this you would want to read every allocated memory location periodically, 'washing it clean' so the accumulated errors wouldn't become double-bit errors (unrecoverable).

I'd put a wash routine in the background process, where it would string-move a block of memory to nowhere in a round-robin way. Not a terrible hit on the cache; we're idle when in the background task so not impacting the most used code. Some latency issue with interrupts and the like.

> I wouldn't want unreliably hardware for any task, even as simple as gaming.

Would you prefer frame-perfect rendering to increased performance?

In my case I've got a large amount of ram (128GB) and run many virtual machines for various purposes, along with that I've got 24TB of drives hooked up with ZFS running to back up those virtual machines, family photos and movies, etc. Being able to know that an error has happened and it's been corrected or handled appropriately (even killing the system and needing a reboot is appropriate) so that data isn't destroyed is a good thing for me.

We just set up a Kubernetes cluster for building C++ and Java software. Thing is - you want to be absolutely sure that the software you software is built correctly and don't have any bit-flips ending up in your final builds, so ECC is an absolute must. Threadripper supporting this allowed us to create that cluster with cheap commodity hardware and made self-hosting the build-farm the clear financial winner, especially since we have tons of free rack-space (came with the building that was bought), already host servers locally - which means a lot support infrastructure is already in place, and have a solar surplus throughout the year in this building.

Ditto on the affordable option for ECC ram. I've also gone the Threadripper route and it is just fantastic for compile times of large systems.

Yea I've had it about 3 weeks now, and I'm still amazed at how much power it's got. I've never had a system where I have so much spare cpu and ram that I can run anything without ever impacting something else. I've actually hit the point where compile times are significantly more disk bound than cpu bound. I'm actually working on a project to take advantage of all this cpu power to help the Perl community a bit by using the large 30k+ libraries and their test suites to validate in-development versions of perl being properly backwards compatible (been a bit of a sticking point in this latest cycle of perl 5 development).

1. desktop or laptop? 2. does it get hot under load?

Reply on news.ycombinator.com