Anybody planning a Ryzen build?

[Pauven]

Got some very interesting results testing different VM's.

 

Installed a Win10 VM from scratch, fully updated it, ran great the whole time.

 

Then installed an Ubuntu 17.04 beta2 VM from scratch, it installed and ran great too.

 

Re-downloaded Lubuntu 16.10, attempted to install a VM from scratch, and almost immediately locked up the server, same as the 3 previous attempts, except this time it locked up so fast that I didn't get the error messages in the log.

 

Seems there is something incompatible with Lubuntu 16.10 on Ryzen.  Possibly the older 4.8 kernel.  Simple enough, I just won't use Lubuntu 16.10.

 

Earlier I installed Ubuntu 16.10 on Ryzen bare metal without any issues.  Since Lubuntu is just a lightweight version of Ubuntu, I figured it would work fine, even with the older kernel.  Not sure if it is having issues only inside of a VM, or something specific to Lubuntu.  Either way, weird.  Lubuntu installed and ran great on my Celeron G1610.  I was trying to use Lubuntu since it seemed to handle running nested VM's better than Ubuntu (I'm trying to run Windows Server 2008 R2 VMWare image in VMPlayer inside Lubuntu).

 

Now that I've figured out these crashes are being caused by Lubuntu, the only remaining unRAID issue is the hanging console.  I'll get the server booted back up and see how long it can go without Lubuntu's help.

 

-Paul

[JIM2003]

Updating you guys on my situation

I took Pauven's suggestion and disabled Global C State control. On top of that, I installed the latest version of Unraid, and flashed ASRock's new BIOS with the microcode update. My system has been up for 36 Hours now, almost 12 hours longer than the previous record. Hopefully some combination of those changes actually fixed it. Crossing my fingers!

[chadjj]

On 4/3/2017 at 9:46 PM, JIM2003 said:

Updating you guys on my situation

I took Pauven's suggestion and disabled Global C State control. On top of that, I installed the latest version of Unraid, and flashed ASRock's new BIOS with the microcode update. My system has been up for 36 Hours now, almost 12 hours longer than the previous record. Hopefully some combination of those changes actually fixed it. Crossing my fingers!

 

Hi Jim2003,

 

How is your up time a couple days later?

 

Chad

[Pauven]

My up-time has just passed 48 hours with Global C-states Control disabled, and still going strong.  Only negatives so far has been a higher idle wattage and temps.

 

During these 48 hours I've only used the server for file storage.  No dockers or VM's.

 

The recent crashing issues I had (with C-states disabled) seem to have been caused by the Lubuntu 16.10 VM, something that didn't occur with Win10 and Ubuntu 17.04 VM's.

 

-Paul

[chadjj]

On 4/3/2017 at 1:44 PM, Pauven said:

Just tried to install Lubuntu in a VM again, and with the server having only been up <5 minutes the install process got much further.  I actually thought it was going to succeed. Eventually I got an error message on the Lubuntu desktop that the "Installer crashed", and then VNC became hung.

 

Then the unRAID log showed stalls were detected on the CPU's and started outputting the Call Traces, same error messages as before.

 

This time the server crashed pretty quickly, within a minute or two.  

 

I can't rule out a defective install ISO, so I'll try with a Windows VM next.

 

-Paul

 

Hey Paul,

 

I did some digging today on when we can expect our AGESA CPU update for our boards and that should come around April 11th.  I found an article today discussing it.  It's in Spanish but Google translate works fine.  Earlier you indicated things being related to C-State but this is one more potential issue being resolved.  Fingers crossed.

 

https://translate.google.ca/translate?hl=en&sl=es&u=http://www.hd-tecnologia.com/confirmada-nueva-bios-con-agesa-1004-para-los-ryzen-5-ryzen-7-aun-no-tienen-esta-actualizacion/&prev=search

 

Chad

[Pauven]

2 hours ago, chadjj said:

Hey Paul,

 

I did some digging today on when we can expect our AGESA CPU update for our boards and that should come around April 11th.  I found an article today discussing it.  It's in Spanish but Google translate works fine.  Earlier you indicated things being related to C-State but this is one more potential issue being resolved.  Fingers crossed.

 

https://translate.google.ca/translate?hl=en&sl=es&u=http://www.hd-tecnologia.com/confirmada-nueva-bios-con-agesa-1004-para-los-ryzen-5-ryzen-7-aun-no-tienen-esta-actualizacion/&prev=search

 

Chad

 

I think that is the first time I've seen a date specified, thanks.  Of course, it may take a while longer for MB makers to provide a new BIOS incorporating AGESA 1.0.0.4.  I doubt we'll actually get a BIOS with AGESA 1.0.0.4 code on April 11th, might take a few more weeks.  Of course, it's certainly possible for some manufacturers to have already updated their BIOS with the new AGESA code release.  ASRock certainly hasn't for my motherboard.

 

As a reminder, here's what AMD says will be in AGESA 1.0.0.4:

 

Quote

We will soon be distributing AGESA point release 1.0.0.4 to our motherboard partners. We expect BIOSes based on this AGESA to start hitting the public in early April, though specific dates will depend on the schedules and QA practices of your motherboard vendor.

 

BIOSes based on this new code will have four important improvements for you

1. We have reduced DRAM latency by approximately 6ns. This can result in higher performance for latency-sensitive applications.
2. We resolved a condition where an unusual FMA3 code sequence could cause a system hang.
3. We resolved the “overclock sleep bug” where an incorrect CPU frequency could be reported after resuming from S3 sleep.
4. AMD Ryzen™ Master no longer requires the High-Precision Event Timer (HPET).

 

 

I don't see how any of that will fix the issue I encountered with the server crashing while installing a Lubuntu 16.10 VM.  

 

In fact, of the 4 items AMD is improving, only the reduced DRAM latency will impact me.  I'm not using programs that do FMA3 operations, I'm not overclocking so I don't encounter the sleep bug, and since Ryzen Master is Windows only, it doesn't affect unRAID (plus, it doesn't even seem to work on Windows VM's).

 

-Paul

[chadjj]

1 hour ago, Pauven said:

 

I think that is the first time I've seen a date specified, thanks.  Of course, it may take a while longer for MB makers to provide a new BIOS incorporating AGESA 1.0.0.4.  I doubt we'll actually get a BIOS with AGESA 1.0.0.4 code on April 11th, might take a few more weeks.  Of course, it's certainly possible for some manufacturers to have already updated their BIOS with the new AGESA code release.  ASRock certainly hasn't for my motherboard.

 

As a reminder, here's what AMD says will be in AGESA 1.0.0.4:

 

 

I don't see how any of that will fix the issue I encountered with the server crashing while installing a Lubuntu 16.10 VM.  

 

In fact, of the 4 items AMD is improving, only the reduced DRAM latency will impact me.  I'm not using programs that do FMA3 operations, I'm not overclocking so I don't encounter the sleep bug, and since Ryzen Master is Windows only, it doesn't affect unRAID (plus, it doesn't even seem to work on Windows VM's).

 

-Paul

 

Hi Paul,

 

With AGESA being addressed it's at least one less variable to factor.  Agreed it may take MB manufacturers a while to get their releases out but by the way the article reads AMD has already provided the code and boards should have an April 11 BIOS update.  We will see...

 

Possibly updating unRAIDs Linux Kernel to 4.10.8 or 4.11rc5 would be an interesting test for the LimeTech team.  Challenge is their NAS stability philosophy, which makes sense to be on the most reliable release.  Experimental or early releases are not really an area of focus but could resolve our Ryzen issues.  Prioritizing resources on Ryzen may not be a high priority as well seeing that AMD is in the minority of market share.  A post was made earlier about having 2 streams but that would be difficult to manage.

 

Quick link to Linux Kernels

https://www.kernel.org/

 

Chad

[Pauven]

On 4/5/2017 at 4:18 PM, Pauven said:

My up-time has just passed 48 hours with Global C-states Control disabled, and still going strong.  Only negatives so far has been a higher idle wattage and temps.

 

During these 48 hours I've only used the server for file storage.  No dockers or VM's.

 

The recent crashing issues I had (with C-states disabled) seem to have been caused by the Lubuntu 16.10 VM, something that didn't occur with Win10 and Ubuntu 17.04 VM's.

 

-Paul

 

Quick update:  I'm approaching 100 hours of up-time (4+ days).  Usage model hasn't changed, still no docker or VM use.

 

Considering it was typically crashing in 2 hours prior to disabling Global C-state Control, this seems to be a winner for me.

 

-Paul

[Pauven]

Just checked ASRock's website, and they just released BIOS v2.0 for my motherboard.  Here's the description.

 

Quote

1.Update AGESA code. (Enhance for Ryzen 5) 2.Precision Boost - Adjusts clock speeds in 25MHz increments in BIOS setup.

 

So it looks like AGESA 1.0.0.4 made it before April 11th.  Color me surprised!

 

I'm going to hold off for now.  I'd rather let the system run as-is to monitor stability.

 

-Paul

[chadjj]

20 hours ago, Pauven said:

Just checked ASRock's website, and they just released BIOS v2.0 for my motherboard.  Here's the description.

 

 

So it looks like AGESA 1.0.0.4 made it before April 11th.  Color me surprised!

 

I'm going to hold off for now.  I'd rather let the system run as-is to monitor stability.

 

-Paul

 

Just checked the ASUS site and no update on my board yet.  Looks like ASRock is ahead of the game.

 

100 hours of uptime is very positive.  Are you planning on adding 1 docker at a time?

 

Chad

Edited April 8, 2017 by chadjj

[coppit]

Quote

 

My gut is telling me that C6 Mode or Global C-state Control is where the magic is hiding. 

My up-time has just passed 48 hours with Global C-states Control disabled, and still going strong. 

 

I know it's a weak data point, but at work, with Intel CPUs and Xen, we had to disable C-states to work around some lockups.

[JIM2003]

On 4/5/2017 at 10:32 AM, chadjj said:

 

Hi Jim2003,

 

How is your up time a couple days later?

 

Chad

I am just about to hit 7 days of up time!

[chadjj]

31 minutes ago, JIM2003 said:

I am just about to hit 7 days of up time!

 

Excellent.  Hopefully I see positive results once my microcode update hits this week.

 

Chad

Edited April 9, 2017 by chadjj

[Dazing]

Can someone give a short overview/summary on Ryzen/UnRaid compatibility?

 

More specific do the R7 CPU's have PCI-e passthrough and are planned to be fully supported?

[ufopinball]

1 hour ago, Dazing said:

Can someone give a short overview/summary on Ryzen/UnRaid compatibility?

 

More specific do the R7 CPU's have PCI-e passthrough and are planned to be fully supported?

 

Short answer, yes.  Some additional details I posted earlier in the thread, but there is a lot more if you go through the rest of the posts that are here.  Trust me, it's worth doing.

 

There are also some drawbacks that will hopefully be ironed out over time.  Ultimately, you may be able to get a better answer if you provide more detailed specifics on what you plan to do.

 

- Bill

[ufopinball]

Quick update from me.  I upgraded to unRAID 6.3.3, and they released a newer BIOS (0515) for the ASUS Prime X370-PRO.  I've been running on both for the past week or so with good results.  I've been paying less attention to uptime figures, and more on just getting my Windows 10 VM configured with all the basic software that I need.  My best uptime was

2 days, 17 hours, 34 minutes, if it matters.

 

After going through the BIOS, I finally found the "Global C-state Control" setting.  I changed it from "Auto" to "Disabled".  Since I still don't have access to sensors (temperatures or fan speeds), the best I can measure is wattage used at the UPS.  With C-State set to "Auto", idle would show as low as 54 watts.  With C-State set to "Disabled", this value did not change ... which was unexpected?

 

So, I turned off the Windows VM and went to bed.  By morning (6 hours or so), my system had hung.  Based on the power readings, I'm guessing the C-State control isn't fully implemented in the BIOS yet.

 

Anyway, just a data point.  I rebooted, left the C-State value as "Disabled" and restarted the Windows 10 VM.  Presumably, things will continue to function normally ... I'll post an update if it doesn't.  Outside of any other factors, I plan to leave this running until there's a new OS or BIOS to try out.

 

- Bill

 

[Dazing]

2 hours ago, ufopinball said:

 

There are also some drawbacks that will hopefully be ironed out over time.  Ultimately, you may be able to get a better answer if you provide more detailed specifics on what you plan to do

 

 

Long term i would like to create a UnRaid box that does:

1. Runs a Windows VM that can render videos and other creative cloud stuff along side maybe some game streaming if that shit actually starts working properly.
2. Runs dockers with Gitlab/Plex/Owncloud etc
3. Docker or Linux VM as web server for development
4. Traditional NAS (doh)

And #1 requires GPU pass though to be effective.

[ufopinball]

4 hours ago, Dazing said:

 

Long term i would like to create a UnRaid box that does:

1. Runs a Windows VM that can render videos and other creative cloud stuff along side maybe some game streaming if that shit actually starts working properly.
2. Runs dockers with Gitlab/Plex/Owncloud etc
3. Docker or Linux VM as web server for development
4. Traditional NAS (doh)

And #1 requires GPU pass though to be effective.

 

Well, there are some more substantial details on my setup, in this post.

 

1)  I am running a Windows 10 VM with GPU pass through.  This is an Asus Radeon 6450 1GB (Desktop) Graphics Card, a non-gaming GPU.  I can rip DVDs or BluRays, and run HandBrake with no problem.  I have assigned 8 cores to my VM so it's got some muscle.  Dunno what you mean by "creative cloud stuff", but it runs things like PhotoShop or Visual Studio.  I'm not really a gamer, so I can't comment on that.  I also only have one GPU, I have never tried to do multiple GPU cards.

 

1a)  You can run Windows 10 or Windows 8.x VMs with the GPU pass through.  Unfortunately, Windows 7 VMs do not fully support this feature.

 

2)  Yes, I run dockers like Plex or Crashplan.  Haven't spent a ton of time on Dockers, but this list could certainly grow and I see no reason why things wouldn't work.

 

3)  Yes, I have a running LAMP VM.  You can visit www.ufopinball.com if you like.  My base LAMP OS is CentOS7, but I imagine this is mainly up to personal preference.

 

4)  Yes, just the standard unRAID setup though.  Details are in my signature, or you can see the previously referenced post.  I have a handful of plugins, the only real issue I have is the sensor drivers (CPU temp, fan speed) aren't included in the current Linux kernel.

 

What sort of Ryzen server were you planning to build?  Specs?

 

- Bill

[chadjj]

9 hours ago, ufopinball said:

Quick update from me.  I upgraded to unRAID 6.3.3, and they released a newer BIOS (0515) for the ASUS Prime X370-PRO.  I've been running on both for the past week or so with good results.  I've been paying less attention to uptime figures, and more on just getting my Windows 10 VM configured with all the basic software that I need.  My best uptime was

2 days, 17 hours, 34 minutes, if it matters.

 

After going through the BIOS, I finally found the "Global C-state Control" setting.  I changed it from "Auto" to "Disabled".  Since I still don't have access to sensors (temperatures or fan speeds), the best I can measure is wattage used at the UPS.  With C-State set to "Auto", idle would show as low as 54 watts.  With C-State set to "Disabled", this value did not change ... which was unexpected?

 

So, I turned off the Windows VM and went to bed.  By morning (6 hours or so), my system had hung.  Based on the power readings, I'm guessing the C-State control isn't fully implemented in the BIOS yet.

 

Anyway, just a data point.  I rebooted, left the C-State value as "Disabled" and restarted the Windows 10 VM.  Presumably, things will continue to function normally ... I'll post an update if it doesn't.  Outside of any other factors, I plan to leave this running until there's a new OS or BIOS to try out.

 

- Bill

 

 

Hi Bill,

 

I have a very similar setup however haven't made it past 8 hrs of up time.  I have global C-State disabled no over clocking straight stock and system hangs.  Also on unRAID 6.3.3.

 

ASUS X370-Pro, Corsair 16GB kit 3000, same one AMD sent out for press release which is a differentiating factor. Ran memtest86 with no errors.

 

I am running several dockers so they may be triggering something different.  Plex (official), ruTorrent (binhex-rtorrentvpn), SickRage (binhex-sickrage), CouchPotato (binhex-couchpotato-git), PlexPy (linuxserver) also tried Deluge in the past but not on Ryzen.

 

If your up time is solid may be docker related.  This would be a new issue based on past results from everyone.  I am looking forward to AGESA this week and will report back results then.

 

-Chad

Edited April 10, 2017 by chadjj

[ufopinball]

4 hours ago, chadjj said:

 

Hi Bill,

 

I have a very similar setup however haven't made it past 8 hrs of up time.  I have global C-State disabled no over clocking straight stock and system hangs.  Also on unRAID 6.3.3.

 

ASUS X370-Pro, Corsair 16GB kit 3000, same one AMD sent out for press release which is a differentiating factor. Ran memtest86 with no errors.

 

I am running several dockers so they may be triggering something different.  Plex (official), ruTorrent (binhex-rtorrentvpn), SickRage (binhex-sickrage), CouchPotato (binhex-couchpotato-git), PlexPy (linuxserver) also tried Deluge in the past but not on Ryzen.

 

If your up time is solid may be docker related.  This would be a new issue based on past results from everyone.  I am looking forward to AGESA this week and will report back results then.

 

-Chad

 

Disabling "Global C-State" on the ASUS Prime X370-PRO seemed to cause problems for me.  I left mine disabled, rebooted, and loaded the Windows 10 VM only to continue to have stability issues.  In other words, disabling made things worse, not better.

 

Right now, I'm back to having the "Global C-State" set to "Auto" (default) and keep the Windows 10 VM running at all times.  The theory is this keeps the processor busy enough to not drop into an idle state that causes instability issues in unRAID.

 

Now, you DO mention running a number of dockers.  In particular, I would guess that ruTorrent should keep the system reasonably busy, assuming you are seeding a few torrents?

 

If nothing else, try setting up the Windows VM and see if things get any better.  I was seeing uptimes in the 5+ day range before I had to reboot for other reasons.

 

If you're still seeing system hangs, try to gauge your uptime duration so we can have it on record.

 

- Bill

 

PS:  As noted earlier, it IS really worth reading through the entire thread.  That way, we aren't repeating information and details that are otherwise known to the other readers of this thread.

 

 

Edited April 10, 2017 by ufopinball

[Dazing]

9 hours ago, ufopinball said:

What sort of Ryzen server were you planning to build?  Specs?

 

- Bill

 

Looking at a 1700 or 1700x, im guessing the x370 chipset is the goto one, 32GB of DDR4 and use my old (but un-used) WD 2TB Red's, 2 256GB SSD's for cache and a GTX 980Ti.

 

From what I've read SLI is the only difference on the MB chipsets. And the only thing that seperates the x model cpu's have turbo boost. Correct?

[Dazing]

23 hours ago, ufopinball said:

Some additional details I posted earlier in the thread, but there is a lot more if you go through the rest of the posts that are here.  Trust me, it's worth doing.

 

I took your advice and skimmed the whole thing, interesting stuff. It led me quite frankly to the conclusion that it a fair amount of time and effort left before its mature enough to be a safe bet, even thought its looking better.

 

Like many others i would also chip in and thank you folks for trying so hard to make it work!

[Pauven]

23 hours ago, ufopinball said:

 

Quick update from me.  I upgraded to unRAID 6.3.3, and they released a newer BIOS (0515) for the ASUS Prime X370-PRO.  I've been running on both for the past week or so with good results.  I've been paying less attention to uptime figures, and more on just getting my Windows 10 VM configured with all the basic software that I need.  My best uptime was

2 days, 17 hours, 34 minutes, if it matters.

 

After going through the BIOS, I finally found the "Global C-state Control" setting.  I changed it from "Auto" to "Disabled".  Since I still don't have access to sensors (temperatures or fan speeds), the best I can measure is wattage used at the UPS.  With C-State set to "Auto", idle would show as low as 54 watts.  With C-State set to "Disabled", this value did not change ... which was unexpected?

 

So, I turned off the Windows VM and went to bed.  By morning (6 hours or so), my system had hung.  Based on the power readings, I'm guessing the C-State control isn't fully implemented in the BIOS yet.

 

 

14 hours ago, chadjj said:

I have a very similar setup however haven't made it past 8 hrs of up time.  I have global C-State disabled no over clocking straight stock and system hangs.  Also on unRAID 6.3.3.

 

ASUS X370-Pro, Corsair 16GB kit 3000, same one AMD sent out for press release which is a differentiating factor.

 

 

It sounds to me that the ASUS X370-Pro, that both of you (Bill and Chad) are using, isn't actually disabling C-states through the BIOS.  There was a repeatable and very noticeable impact to me disabling the C-states on my ASRock, as idle power consumption increased anywhere from 7 to 27 watts, and CPU temperatures also idled higher (hard to say how much, but seems 5-10c higher).  Both of which makes perfect sense considering I'm basically turning off all the CPU power saving modes by disabling C-states.

 

It stands to reason that if the BIOS switch isn't working, then the resulting fix isn't actually being applied, so you're still having crashes.

 

I think it would still be helpful to determine for absolute certainty that Global C-states Control is not being disabled.  There should be a way to check this in Linux, but in my searching I could not find a way to do it for Linux 4.x and AMD CPU's (most of the commands I found predated the 2.6 Kernel).

 

It might be easier to test in Windows (bare metal, not VM).  I'm sure there's a Windows diagnostic program that can show what C-state the CPU is using at any given time, perhaps even the Ryzen Master program itself (I haven't played with it myself so I don't know).  

 

If you can show that the C-state switch is broken, it will help convince ASUS to fix the BIOS when you open a ticket, especially if multiple users are opening tickets for the same problem.

 

On the other hand, it is still possible that my motherboard is not behaving correctly when I disable Global C-state Control, and perhaps the real fix lies elsewhere.  I know that my motherboard is not the only one for which this fix is working, but I think the other MB's may have been ASRock too.

 

-Paul

[Pauven]

16 hours ago, ufopinball said:

3)  Yes, I have a running LAMP VM.  You can visit www.ufopinball.com if you like.  My base LAMP OS is CentOS7, but I imagine this is mainly up to personal preference.

 

Nice to see a fellow pinball enthusiast in the neighborhood.  I'd been meaning to ask about your username, but I just figured you were a fan of Attack from Mars.  I think I may have seen your handiwork on a AfM machine here in Atlanta.

 

I have a long-time business partner that is a pinball freak.  So much so that he convinced me we should build pinball machines as our next career.  To that extent, I actually built a pinball machine from scratch (I bought a CPR playfield and various mechanical assemblies, but the rest is mine).  It took years of research and development, and was a lot more complex than I ever anticipated.  I detailed it all here:

 

http://pinballchameleon.blogspot.com/

 

I took my completed "Modern Firepower" to the Southern Fried Gameroom Expo a few years ago, and it was a smash hit.  The response was enough to prompt me to keep up development of the platform.

 

I haven't updated the site in a few years, but I'm actually quite a bit further along in development of a 2nd generation machine.  My software is much more capable now, as I keep adding new features, improving performance, and fixing the odd bug.  Plus, I've also developed my own next-generation circuit boards to interface the PC to the mechanics.  Basically, what I've developed is a plug and play kit to build your own pinball machine, a Pinball Construction Kit (too bad the name was wasted on a video game).  My kit could be used by a weekend hobbyist, or even a small manufacturer that wants to build pinball machines  - they could focus on creative efforts rather than engineering which I've already done for them.  It's basically the easy DIY kit I wanted 8 years ago when I started this project.

 

There's a ton of stuff I haven't revealed on the website, mainly because I'm afraid someone would beat me to market (assuming there is a market).  Being a one-man do-it-all shop, it's been very slow going.  I have a lot of very unique solutions I haven't seen anywhere else.  I think a lot of my solutions are unique simply because I know nothing about pinball machines (never owned a real one before I made my own), so I was pretty much starting from scratch on how everything worked, and naturally came up with different solutions to many problems.

 

I haven't sold anything yet, still in the development and testing phase.  There's a local guy that has built some custom machines (and wrote a lot of code for the Pinball Circus project, if you know what that is) and he's going to build a machine around my kit once I get it to him.  My business partner and I have also talked about making a run of finished machines, and we've got a few theme ideas, but I've got to get my kit complete before I can think about making full machines.

 

Anyway, I guess this is pretty off-topic here, sorry.  Just couldn't resist the shout out after I saw your website.

 

-Paul

[ufopinball]

5 hours ago, Pauven said:

 

Nice to see a fellow pinball enthusiast in the neighborhood.  I'd been meaning to ask about your username, but I just figured you were a fan of Attack from Mars.  I think I may have seen your handiwork on a AfM machine here in Atlanta.

[...]

I haven't sold anything yet, still in the development and testing phase.  There's a local guy that has built some custom machines (and wrote a lot of code for the Pinball Circus project, if you know what that is) and he's going to build a machine around my kit once I get it to him.  My business partner and I have also talked about making a run of finished machines, and we've got a few theme ideas, but I've got to get my kit complete before I can think about making full machines.

 

Anyway, I guess this is pretty off-topic here, sorry.  Just couldn't resist the shout out after I saw your website.

 

-Paul

 

Yup, the AFM mini-saucer kit is mine, glad to see it's still making the rounds.  Your Modern Firepower looks to be a pretty involved project!!  I wish I could full screen the video, but the basics that I saw were quite impressive.  Personally, I have only done mods, I haven't tried to reprogram a game yet.  I do have a Cactus Canyon Continued build, but it was a by-the-numbers sort of thing based on the website of the guy who created it.

 

For "Pinball Circus", you mean the game from Python?  I've played the one at the Pinball Hall of Fame in Las Vegas.  

 

Back on topic, it looks like ASUS has released a new BIOS (0604) just today, and it includes the new AMD "Update AGESA to 1.0.0.4a" code.

 

For the moment, my setup has been running smoothly since the last reboot ... er, yesterday.  Been mainly focusing on being productive and getting things done on the machine, rather than tweaking the machine itself.  I'll find some relatively convenient time to upgrade my BIOS later today and hope things continue to stay solid.

 

Otherwise, I guess it's up to anyone else if they wanna take the plunge with us, or if they feel better waiting for more of the bugs to be ironed out.  I'm happy enough with the stability of my rig, and hope that things will continue to get better.  With the pre-latest BIOS (0515) my system will idle as low as 43c while in the Setup menus.  Far better than the 60c that it was reporting with I think 0504.

 

- Bill