| 09:01 | Exaga> | jnettlet: after ~24 hours of Clearfog experience under my belt, I can say it's a very neat piece of technology indeed. Only concern I have with it is the heat produced by the SoC. It's 62 degrees Celsius just idling. |
| 09:01 | Exaga> | If i start compiling on it I can imagine there will be trouble :> |
| 09:01 | jnettlet> | Exaga: unfortunately yes. The cpu-idle and frequency scaling support does not function properly on the chip. |
| 09:02 | jnettlet> | nah. it may go up by 10C |
| 09:02 | Exaga> | haha and then some :> |
| 09:02 | jnettlet> | it just doesn't go down :( |
| 09:02 | Exaga> | i usually hammer all cores "make -j " style |
| 09:03 | jnettlet> | the heat seems to be a byproduct of the factory silicon process. The iMX8M runs much warmer than expected as well....same process. |
| 09:03 | jnettlet> | the iMX8M-mini on the newer process is back down around where the iMX6 was in terms of thermals |
| 09:04 | Exaga> | although it's a SBC it is a router after all |
| 09:04 | Exaga> | ahhh i see |
| 09:04 | Exaga> | well it's going to be used as a router, running debian |
| 09:04 | jnettlet> | if you enable cpuidle support the power consumption actually increases....OUCH! |
| 09:05 | Exaga> | somewhat glorified and deluxe router :D |
| 09:05 | jnettlet> | they are a beast for a router. I run two at my home, one as a bridge to get around the stupid fiber gateway my ISP runs, and the other to separate off my office network from my home network |
| 09:05 | Exaga> | very cool |
| 09:06 | Exaga> | i haven't seen any other device which is as perfect for the job as the clearfog |
| 09:06 | jnettlet> | I don't know where it is with Debian, probably things you would have to do by hand. But for OpenWRT it can easily handle SQM QOS to alleviate bufferbloat over your internet connection. |
| 09:06 | Exaga> | it's like a geek's wildest fantasies come tru lol |
| 09:07 | Exaga> | I will look into OpenWRT |
| 09:07 | jnettlet> | https://photos.app.goo.gl/ZNFSabfQbZFc7UMJA |
| 09:08 | Exaga> | wow thats pretty cool |
| 09:08 | jnettlet> | My connection is a 300/300 fiber. I dedicate 150/150 to my office and 100/100 to my home. The other 50/50 is "shared" bandwidth that QOS favours to my office. But if nothing is going on there then my home network will boost to 150/150 |
| 09:09 | Exaga> | i hate to admit this... but internet speed is only relevant to me for playing games |
| 09:09 | Exaga> | otherwise it's like the size of your car engine |
| 09:09 | Exaga> | gets you there, eventually |
| 09:10 | jnettlet> | I am constantly moving around big images, and pulling down source, updating test systems. It saves me lots of idle time |
| 09:10 | jnettlet> | but for games the SQM QOS is what you need. guaranteed lowered latency for all packets, even if a big download is happening simulataneously |
| 09:11 | Exaga> | noted. thanks for the info |
| 09:12 | jnettlet> | anytime. enjoy playing around and you know where to ask if you have any questions |
| 09:12 | Exaga> | sure thing jnettlet. many thanks once again :> |
| 09:13 | Exaga> | rabeeh still lon holiday? lol |
| 09:19 | jnettlet> | nah, just busy with the new designs |
| 09:20 | jnettlet> | SolidRun is a busy place these days. Every time I make it down to the office there are new faces |
| 09:21 | jnettlet> | I am excited to finally get some benchmarks of the new Workstation posted soon. 16-cores at 2.4ghz is really quite beastly |
| 09:22 | Exaga> | 16 cores? LOL??? What are you guys planning at SolidRun? hahahaha |
| 09:26 | jnettlet> | Building an ARM workstation for developers. Not to mention lots of edge computing |
| 09:32 | Exaga> | ARM <3 now you have my undivided attention lol |
| 09:32 | Exaga> | what kind of arm workstation? |
| 09:32 | Exaga> | development in what area? |
| 09:33 | Exaga> | is this workstation for in-house or will it be available to Joe Public? |
| 09:36 | jnettlet> | Nope it is for community developers. Not cheap, but much less than anything else out there. |
| 09:37 | jnettlet> | https://www.solid-run.com/nxp-lx2160a-family/clearfog-itx/ |
| 09:37 | Exaga> | i can imagine 16-cores at 2.4ghz would be rather costly |
| 09:39 | jnettlet> | well the pre-sale price is $550 for the COM Express module and the carrier. That is 1 board per developer. It goes up once we hit production. |
| 09:39 | jnettlet> | That is without memory. |
| 09:40 | Exaga> | give me a moment or three until I stop drooling over my keyboard, and your website |
| 09:40 | jnettlet> | With a decent GPU, 32GB of memory, nvme, case power supply. The lowest I have spec'd out is about $1000, but with some nicer upgrades I see an average price of around $1300 |
| 09:41 | jnettlet> | This is the case we are using internally. https://www.amazon.de/CA-H200W-BR-Mini-ITX-PC-Geh%C3%A4use-Schwarz-Device/dp/B0776RK46S/ref=sr_1_6?keywords=nzxt+h200i&qid=1559202048&s=gateway&sr=8-6 |
| 09:41 | jnettlet> | you can get it cheaper without the LEDS and intelligent fan controller |
| 09:42 | jnettlet> | not too big, but not enormous and can take 4 2.5" drives so matches the number of SATA ports on the board |
| 09:43 | Exaga> | "Push the limit of the network’s edge with CEx7 LX2160A" = I think you will be breaking those limits very soon with this device |
| 09:43 | Exaga> | right now I'm thinking how i can generate that kind of cash to buy one :D |
| 09:44 | Exaga> | OMG! (x 1000) |
| 09:44 | Exaga> | this technology is seriously way ahead of anything else |
| 09:45 | jnettlet> | and the entire package at full power is only 32W |
| 09:46 | jnettlet> | With a Radeon RX 550 Pulse, you can hit a full workstation running at < 100Watts |
| 09:46 | Exaga> | 32Watts? HUH??? |
| 09:47 | jnettlet> | yep it is a beast. |
| 09:49 | Exaga> | I think you guys might be getting arrested for making hardware like this. It's got to bne illegal surely? LOL |
| 09:49 | Exaga> | be* |
| 09:49 | Exaga> | thanks for letting me know about this jnettlet |
| 09:49 | Exaga> | a large portion of my day will be spent investigating this device |
| 09:50 | Exaga> | imagine what could be achieved with this. it's mind-blowing |
| 09:50 | Exaga> | what am i saying? you're the guy working on it. DOH! :D |
| 09:52 | jnettlet> | thanks. We think with the price point we can finally get full ARM workstations out to the developer community. No more x86 on the desktop developing ARM code |
| 09:54 | Exaga> | ok. in terms of supplying this and aiming it at the ARM developers community... |
| 09:55 | Exaga> | my first impression is that it's the be-all and end-all of bona fide ARM technology |
| 09:55 | Exaga> | the price... OUCH! but then this kind of technology does not come cheap |
| 09:56 | Exaga> | i can imagine quite a few of my friends and colleagues will be struck dumb when they see this |
| 09:57 | Exaga> | I can imagine, as it will be for the more serious (extremely serious) ARM developer, it's out of the price-range of most "hobbyist" users such as myself |
| 10:01 | Exaga> | jnettlet: every once in a while new technology arrives on the scene that canges the way we think, work, and operate. The Clearfog ITX is such technology. |
| 10:02 | Exaga> | You guys at SolidRun seem to have a knack for innivation |
| 10:02 | Exaga> | innovation* |
| 10:03 | jnettlet> | Exaga: thanks for the support. We hope so. The price is definitely much more than a normal ARM SBC. But compared to the other ARM workstations out there, we are still 50% of the price...at least |
| 10:03 | jnettlet> | sometimes 4x cheaper |
| 10:03 | jnettlet> | It is a passion |
| 10:05 | Exaga> | well, show me a 16 core 2.4GHz ARM device that supports 64GB DDR4 RAM with a M.2 interface and I will concede :> |
| 10:05 | Exaga> | hahaha |
| 10:05 | Exaga> | except the clearfog ITX i mean |
| 10:06 | jnettlet> | there are a couple. Also the new Power9 based options from Talos are another option architecture |
| 10:09 | Exaga> | ok, just for comparison I will review the Talos II TL1MB1 against the Clearfog ITX |
| 10:09 | Exaga> | it's comparable in price but not technology |
| 10:16 | jnettlet> | I am running the phoronix benchmark suite he just ran on the Talos board last night on the CF-ITX right now |
| 10:19 | Exaga> | jnettlet: ok now you're just showing off! Just tell me how I get a job at SolidRun lol |
| 10:19 | jnettlet> | the board isn't in my hands. This is the current bringup sample that is down in Israel right now |
| 10:20 | Exaga> | who ran the benchmark suite on the CF-ITX? |
| 10:22 | jnettlet> | I am running it |
| 10:23 | jnettlet> | although I need to setup a new Ubuntu chroot for it to be apples to apples comparison |
| 10:23 | jnettlet> | I will be posting my results by the weekend |
| 10:23 | Exaga> | cool |
| 10:23 | Exaga> | sorry, you wrote "I am running the phoronix benchmark suite he just ran on the Talos board last night |
| 10:24 | Exaga> | and i wonderede who "he" was :> |
| 10:24 | jnettlet> | that he is Michael who runs Phoronix |
| 10:24 | Exaga> | ahh ok |
| 10:27 | Exaga> | so the Talos II TL1MB1 isn't ARM therefore it's not something I would be interested in, and as such, doesn't compare to the CF-ITX |
| 10:30 | jnettlet> | well it is an interesting comparison. It is another alternative architecture. It is only 4 cores, but 16-threads...and it has a much higher clock speed |
| 10:30 | Exaga> | and supports 128GB DDR4 RAM |
| 10:30 | jnettlet> | as an alternative to x86_64 I think it is an interesting comparison |
| 10:31 | jnettlet> | yep. We only support 64GB DDR4 |
| 10:31 | Exaga> | from what I read, Power9 is RISC based? |
| 10:31 | jnettlet> | although there is a point when more RAM isn't so useful since you don't have enough cores to process that much data |
| 10:31 | jnettlet> | yes RISC PPC |
| 10:31 | jnettlet> | it is a fully open source hardware design |
| 10:32 | Exaga> | I understand and have just found out how much 64GB DDR4 2666MHz RAM makes a difference to 16GB DDR3 1600MHz RAM |
| 10:32 | Exaga> | i very recently upgraded my Windows PC |
| 10:33 | Exaga> | yes, sorry to admit i do have one |
| 10:33 | Exaga> | for games :P |
| 10:33 | jnettlet> | you don't have to be sorry. The right tool for the job |
| 10:33 | Exaga> | exactly. great that you have that level of understanding :> |
| 10:34 | jnettlet> | Plus we partner with Microsoft, so I had to swallow that pill a long time ago |
| 10:34 | Exaga | 10:34 * Exaga gulps! |
| 10:34 | jnettlet> | and really their IoT team is really focused on OSS, and quite good with collaboration |
| 10:35 | Exaga> | Microsoft didn't get where they are today by being useless, that's for sure. some of their software is pretty good |
| 10:41 | Exaga> | jnettlet: i always assumed you were based in Israel |
| 10:43 | jnettlet> | Exaga: nope. a US Expat living in Denmark |
| 10:43 | Exaga> | ahhh interesting |
| 10:43 | Exaga> | and, if you don't mind me asking, how did you get involved with SolidRun? |
| 10:48 | jnettlet> | OpenSource collaboration efforts. I was working at OLPC at the same time that SolidRun was formed and Rabeeh and Kossay released the first Cubox based on the Marvell Dove chipset. |
| 10:48 | Exaga> | i remember that time :> |
| 10:49 | jnettlet> | At OLPC we were also partnering with Marvell but a different division. They still used the same Graphics chipset, and Rabeeh and I ran into each other in the Community while integrating support |
| 10:52 | Exaga> | the case for the CF-ITX you gave me the link to is very stylish and looks well made, robust |
| 10:52 | Exaga> | a little expensive for what it offers though |
| 10:52 | jnettlet> | well you can cost down. We of course will spend a bit more for our demo models and internal hardware |
| 10:53 | Exaga> | sure. I think it's a great case and very suitable |
| 10:53 | Exaga> | quality is something that SolidRun are not stringent on, from my experience |
| 10:53 | jnettlet> | it also has a nice clear side-panel so you can see the hardware |
| 10:54 | jnettlet> | we try. but nobody is perfect. Not even Apple :) |
| 10:54 | Exaga> | yes for the asethetially-minded i'm sure that's a big plus :D |
| 10:54 | jnettlet> | and they have way more money than we do |
| 10:54 | jnettlet> | and tradeshows |
| 10:55 | Exaga> | bah, if it were Apple it would be manufactured on the dark side of the moon from NASA-grade aluminium |
| 10:55 | Exaga> | and you'd have to sell your house to be able to afford to buy it |