Personally, I think you're right. You can buy blades that have similar hardware to gamer hardware, but it's not the primary market. And look at exactly what happened when they used a consumer market ryzen: they found a problem that it's good to be aware of, that this particular processor/mobo combo did not perform well without fast RAM. We don't know all the details, but certainly the standard market of blade servers supports ECC RAM, which is another category that most gamers aren't dealing with. So yeah, I think there's sense to making machines that resemble consumer products.BlueTemplar wrote: βFri Oct 04, 2019 4:05 pm Guys, they have to test on hardware approaching what the buyers of the game will have... if you start talking about "blades", aren't you going into very specific server hardware ?
Friday Facts #315 - New test servers
Re: Friday Facts #315 - New test servers
-
- Burner Inserter
- Posts: 9
- Joined: Thu Jul 23, 2015 6:01 pm
- Contact:
Re: Friday Facts #315 - New test servers
XKCD just came to my mind since I'm currently reading how toIn real time this is 8 minutes and 50 seconds, a long time for a team member to sit around for results before they can push.
https://www.xkcd.com/303/
-
- Long Handed Inserter
- Posts: 58
- Joined: Fri Mar 09, 2018 7:33 pm
- Contact:
Re: Friday Facts #315 - New test servers
Is it just me or is that train in the video not staying exactly in the center of the screen? Especially when driving horizontally the yellow wagon looks like it's jittering a bit.
- lottery248
- Inserter
- Posts: 21
- Joined: Thu Jan 04, 2018 9:41 am
- Contact:
-
- Burner Inserter
- Posts: 6
- Joined: Thu May 10, 2018 12:47 am
- Contact:
Re: Friday Facts #315 - New test servers
There are a number of benchmark posts showing how RAM makes a big difference in Factorio. If you buy a fancy CPU but don't get the RAM to match, then the RAM becomes the performance bottleneck.
Proof:
https://docs.google.com/spreadsheets/d/ ... 1699517607
https://imgur.com/2FPr3su
https://www.reddit.com/r/factorio/comme ... ed_fpsups/
https://steamcommunity.com/app/427520/d ... 701196558/
A few years ago when I built my new PC I paired these together, and the performance boost was big:
ASUS PRIME Z270-A LGA 1151
G.SKILL TridentZ Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3600 (PC4 28800)
Intel Core i7-7700K Kaby Lake Quad-Core 4.2 GHz
Proof:
https://docs.google.com/spreadsheets/d/ ... 1699517607
https://imgur.com/2FPr3su
https://www.reddit.com/r/factorio/comme ... ed_fpsups/
https://steamcommunity.com/app/427520/d ... 701196558/
A few years ago when I built my new PC I paired these together, and the performance boost was big:
ASUS PRIME Z270-A LGA 1151
G.SKILL TridentZ Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3600 (PC4 28800)
Intel Core i7-7700K Kaby Lake Quad-Core 4.2 GHz
Re: Friday Facts #315 - New test servers
I haven't said anything against backups. You ignored the other things. Explain to me why they need an uninterruptible power supply or a sealed temperature controlled room.DaleStan wrote: βFri Oct 04, 2019 4:05 pm "Not critical" doesn't mean "no backups". This forum isn't critical, but it should still be backed up. mods.factorio.com isn't critical, but it should still be backed up. Your family photos aren't critical, but they should still be backed up.
(This post has been brought to you by the backup software your favorite podcaster, Twitch streamer, or Youtuber recommends.)
And Wube probably is hosting critical things. Even if their source is cloud-hosted, they probably still store authentication secrets, RSA private keys, and similar things locally.
Re: Friday Facts #315 - New test servers
>We knew beforehand that Ryzen liked fast RAM, but we didn't realize how significant of a difference it could make.
wube doesnt read ryzen benchmark threads confirmed
wube doesnt read ryzen benchmark threads confirmed
Re: Friday Facts #315 - New test servers
This is Prague. There is no server market or common set of companies which sell server hardware to buy from All of those machines are just desktop systems we've built from parts purchased online. If I compare the cost of the hardware vs what I know the same performance hardware would cost in blade form we got it for about 1/3rd the cost. It's a cheap game to play.ledow wrote: βFri Oct 04, 2019 4:00 pm Same (but not former).
I'm disappointed that they haven't just got a blade setup for everything. Having "test" machines that are off-the-shelf servers is an expensive game to play, especially as they obsolete.
Buy a blade system - every time things take too long, shove in another blade, or another processor into the blades. If you have any kind of parallelism in the process, and buy decent blades in the first place, it'll walk over i9's in a box. Not to mention can be pulled for other things via virtualisation etc.
I wouldn't be surprised if they haven't spent more, and got less, than just buying a blade system in the first place and using it for everything.
I've used server CPUs and I've used workstation CPUs and workstation CPUs win when it comes to the workload that we have. We've tested ourselves and had people test the same tests on their 48 core server systems with terrible results.
If you want to get ahold of me I'm almost always on Discord.
Re: Friday Facts #315 - New test servers
the gui has been cut outHenry Loenwind wrote: βFri Oct 04, 2019 11:09 pm Is it just me or is that train in the video not staying exactly in the center of the screen? Especially when driving horizontally the yellow wagon looks like it's jittering a bit.
Re: Friday Facts #315 - New test servers
Current systems engineer, and I think wube made the right choice re: test hardware. You have different priorities and make different decisions when you're in the 'S' part of SMB than you do in the 'M' part.
I would never put that stuff in my network, but I have different focuses/needs/requirements for scalability, availability, etc.
So, while I agree that for the test environment, their size, and their probably needs, they made (at least in general) good choices in buying consumer-class hardware, even if I'd have put different components in there.
Now for their non-dev/test environment, you guys should definitely pick up at least one good HPE or Dell server. The reasons would be reliability, and predictive failure. That hardware will fail too, but 'real' servers can often predict a failure or have redundant components, as well has have better security down at the component level. I know for a fact I could get Dell and HPE equipment to my last employer's Roermond office, and if we can get it down there in the South Netherlands, you can get it in Amsterdam.
If I didn't live in the US, I'd come over and help you out for free for a day or three, just as a thank you because I love Factorio so much. Unfortunately, I don't love it THAT much that I'd pay thousands in airfare and lodging to donate a week for free. Wish I could though!
Edit: Kinda makes me sick that we're about to send some older Dell R620s and R720s to the scrapper. I guess it's like when my mother used to tease me about finishing my food because there were starving children in Africa. She stopped when I was mature enough to make the logistical argument that there's no conceivable way they'd benefit, seeing as how the food was here and there is no economical way to get it over there before it spoiled.
Maybe I should just build a rocket and load the servers on it.
I would never put that stuff in my network, but I have different focuses/needs/requirements for scalability, availability, etc.
So, while I agree that for the test environment, their size, and their probably needs, they made (at least in general) good choices in buying consumer-class hardware, even if I'd have put different components in there.
Now for their non-dev/test environment, you guys should definitely pick up at least one good HPE or Dell server. The reasons would be reliability, and predictive failure. That hardware will fail too, but 'real' servers can often predict a failure or have redundant components, as well has have better security down at the component level. I know for a fact I could get Dell and HPE equipment to my last employer's Roermond office, and if we can get it down there in the South Netherlands, you can get it in Amsterdam.
If I didn't live in the US, I'd come over and help you out for free for a day or three, just as a thank you because I love Factorio so much. Unfortunately, I don't love it THAT much that I'd pay thousands in airfare and lodging to donate a week for free. Wish I could though!
Edit: Kinda makes me sick that we're about to send some older Dell R620s and R720s to the scrapper. I guess it's like when my mother used to tease me about finishing my food because there were starving children in Africa. She stopped when I was mature enough to make the logistical argument that there's no conceivable way they'd benefit, seeing as how the food was here and there is no economical way to get it over there before it spoiled.
Maybe I should just build a rocket and load the servers on it.
-
- Fast Inserter
- Posts: 122
- Joined: Fri Jun 17, 2016 8:17 pm
- Contact:
Re: Friday Facts #315 - New test servers
Agreed. I had the same combination "aww, that's so adorable" / "oh dear god why" reaction as I think a number of other people here did, but then I realized that it's entirely possible that each U of my FY19 hardware cost more than their entire setup. For testing and such, the high speed consumer-grade hardware is a nice choice.DevAudio wrote: βSat Oct 05, 2019 2:36 am Current systems engineer, and I think wube made the right choice re: test hardware. You have different priorities and make different decisions when you're in the 'S' part of SMB than you do in the 'M' part.
I would never put that stuff in my network, but I have different focuses/needs/requirements for scalability, availability, etc.
So, while I agree that for the test environment, their size, and their probably needs, they made (at least in general) good choices in buying consumer-class hardware, even if I'd have put different components in there.
Now for their non-dev/test environment, you guys should definitely pick up at least one good HPE or Dell server. The reasons would be reliability, and predictive failure. That hardware will fail too, but 'real' servers can often predict a failure or have redundant components, as well has have better security down at the component level. I know for a fact I could get Dell and HPE equipment to my last employer's Roermond office, and if we can get it down there in the South Netherlands, you can get it in Amsterdam.
Ditto with "use a real server in prod", but I'll disagree on HPE or Dell. If you aren't big enough to really squeeze them, you end up way overpaying from either of those vendors -- for smaller one-time orders, a low overhead whitebox reseller like Thinkmate (or whatever the EU equivalent is) would likely serve them better.
- eradicator
- Smart Inserter
- Posts: 5207
- Joined: Tue Jul 12, 2016 9:03 am
- Contact:
Re: Friday Facts #315 - New test servers
Maybe @Rsed can put them into his luggage on his next flight to europe ;p.
Author of: Belt Planner, Hand Crank Generator, Screenshot Maker, /sudo and more.
Mod support languages: ζ₯ζ¬θͺ, Deutsch, English
My code in the post above is dedicated to the public domain under CC0.
Mod support languages: ζ₯ζ¬θͺ, Deutsch, English
My code in the post above is dedicated to the public domain under CC0.
-
- Filter Inserter
- Posts: 362
- Joined: Mon Aug 01, 2016 2:38 pm
- Contact:
Re: Friday Facts #315 - New test servers
I love the video! The music certainly added to the excitement. He plays like I do, enemies at MAXIMUM OVERLOAD! I enjoy barely hanging on against waves of countless biters. My UPS dies, but the warfare of my factory machine against the biters is endless. Love this game with the mods!
Re: Friday Facts #315 - New test servers
yeah Factorio is a great combat game!RocketManChronicles wrote: βSat Oct 05, 2019 2:19 pm I love the video! The music certainly added to the excitement. He plays like I do, enemies at MAXIMUM OVERLOAD! I enjoy barely hanging on against waves of countless biters. My UPS dies, but the warfare of my factory machine against the biters is endless. Love this game with the mods!
I wish devs would see how much potential it could have...
- BlueTemplar
- Smart Inserter
- Posts: 3170
- Joined: Fri Jun 08, 2018 2:16 pm
- Contact:
Re: Friday Facts #315 - New test servers
Haven't they just done that ?
BobDiggity (mod-scenario-pack)
Re: Friday Facts #315 - New test servers
you mean giving us more and more toys to play arround to make killing biters easier with every update?
Biters need a buff. Not in numbers but in strenght and resistance.
At least buff them so they can keep up with endless damage research and uranium ammo
Re: Friday Facts #315 - New test servers
Damn this Red Alert soundtrack makes me want those self assembling building animations in Factorio that the Command & Conquer games had.
Re: Friday Facts #315 - New test servers
Does the massive ammount of cores in your new CPUs imply, that the game now does significantly benefit from more than four cores?
I would have expected single-thread performance and RAM speed to still be the by far most important factors when it comes to Factorio UPS. Also i would have expected Factorio to be a pretty bad workload for virtualization because of the high RAM throughput demands.
I would have expected single-thread performance and RAM speed to still be the by far most important factors when it comes to Factorio UPS. Also i would have expected Factorio to be a pretty bad workload for virtualization because of the high RAM throughput demands.
-
- Inserter
- Posts: 39
- Joined: Mon May 08, 2017 8:13 pm
- Contact:
Re: Friday Facts #315 - New test servers
That is actually what I hoped they'd do to "discourage" turret creep. Let them have a bit of OOMPH when placed. Let it be folded up like those outrigger leg supports used on heavy vehicles like cranes. Have them appear all folded up, then let us see them unfold their legs from an upright position to down on the ground, then lift itself up to operational height. It takes negligible time when turrets are placed for defense, but they are vulnerable if deployed under fire.
I suppose it could be used for other machinery too, but I want it most for turrets.
Re: Friday Facts #315 - New test servers
Traditional turret creep works by building lots of turrets starting outside the enemy's aggro range where new turrets tend to be less far away from the enemy than older turrets.warlordship wrote: βSun Oct 06, 2019 3:45 am That is actually what I hoped they'd do to "discourage" turret creep. Let them have a bit of OOMPH when placed. Let it be folded up like those outrigger leg supports used on heavy vehicles like cranes. Have them appear all folded up, then let us see them unfold their legs from an upright position to down on the ground, then lift itself up to operational height. It takes negligible time when turrets are placed for defense, but they are vulnerable if deployed under fire.
One common property of turret creep is, that new turrets are protected by old turrets while they are build, so build time does not really matter that much (but squishiness of unfinished turrets does).
Multiple working turret creep patterns have been independently invented and successfully used by lots of players in traditional RTS games like Dune 2, C&C, and Warcraft 2.