|
So where exactly is this antenna going to go?
|
#
?
Apr 17, 2023 01:12
|
|
|
|
| # ? May 15, 2024 13:19 |
|
|
EoRaptor posted:NFC has nothing to do with Bluetooth. I was aware that NFC as in "an RFID tag" had nothing to do with BT, but every implementation of smarter NFC (like phone payments, unlocking your hotel door with your phone, etc., etc.) that I've seen has been explicitly tied to BT, in that if I turn off BT on my phone I'm also turning off NFC capabilities. So I had assumed that "smart" NFC was piggybacking on BT to provide the transmission link... hobbesmaster posted:Also nrf5340s are everywhere ...and this is probably why: "Dual-core Bluetooth 5.3 SoC supporting Bluetooth LE, Bluetooth mesh, NFC, Thread and Zigbee". So, separate protocols entirely, but commonly implemented together in hardware. Cool. Now I know.
|
|
#
?
Apr 17, 2023 01:55
|
|
|
mdxi posted:So, separate protocols entirely, but commonly implemented together in hardware. Cool. Now I know. For readers yes, the ‘tag’ is what would go into the MB somewhere with a simple serial interface to get updated. That’s the benefit here, that all the costs are in a device you already own (smart phone) so all the MB needs to add is a tiny nfc chip that can sit on an already existing serial bus, has some flash rom, and a conveniently located loop of wire.
|
|
#
?
Apr 17, 2023 03:08
|
|
|
a lot (maybe most, even?) motherboards these days that lack a pc speaker instead beep through your on-board garbage ac97 chip, which is very funny because people definitely remember to plug their speakers in when building a pc, and also many new monitors totally don't ship with builtin speakers you're supposed to use through dp/hdmi it's all very stupid because one of those tiny pc speakers adds probably $.03 to the BOM? e: and the mobo mfgs will probably take the reasonable way out of this mess and start shipping bluetooth with their motherboards so you can hear the beeps on your anroid phone
|
|
#
?
Apr 17, 2023 09:16
|
|
|
I just set up my 7800x3d build today and saw that there were headers for PC speakers, but it was a $480 ASUS board so it also had the POST code LED which doubles as CPU temperature display once booted, pretty neat if a ripoff price.
|
|
#
?
Apr 17, 2023 11:37
|
|
|
hobbesmaster posted:So where exactly is this antenna going to go? I/o shield would make sense. The last two boards I purchased (msi) have a built in i/o shield, which is frankly fantastic. You can't lose it, you can't misalign the board with it and get metal tabs bent in front of your ports, and you don't have to snap it in to the case
|
|
#
?
Apr 17, 2023 13:38
|
|
|
hobbesmaster posted:So where exactly is this antenna going to go? on the huge piece of decorative plastic that's over the IO panel!
|
|
#
?
Apr 17, 2023 14:21
|
|
|
So floating above and parallel to the VRMs? This just feels like one of those things where the EEs in layout and EMC ask the product manager how much they really want this.
|
|
#
?
Apr 17, 2023 14:49
|
|
|
Just populate the missing two pins on the speaker header and use the resulting 4 as a combination speaker and 1-wire connector (so you'd get like [+SPKR | DATA | +5V | GND]) so you can plug some weird overpriced daughterboard into it and get the post code and whatever other diagnostics including beeping. Now board vendors can claim they have advanced debugging support and also charge way too much for it without actually doing much of anything, and as a bonus you can detach the thing once it works for the ~clean aesthetic~. Everybody wins, except the consumer, and as a bonus maybe we'd convince board makers to stop removing the speaker header since sometimes hearing the beeps is legit useful. Also a specific indicator of "memory training in progress" while I'm dreaming.
|
|
#
?
Apr 17, 2023 14:58
|
|
|
Just make them not break imo
|
|
#
?
Apr 17, 2023 16:51
|
|
|
Put a BMC on all boards.
|
|
#
?
Apr 17, 2023 16:52
|
|
|
Combat Pretzel posted:Put a BMC on all boards. Intel did that a long time ago, it’s called vPro
|
|
#
?
Apr 17, 2023 17:10
|
|
|
my board has a com port header doing nothing, just send the whole boot log over that like linux does, you could get real detailed, like tell you what dimms were detected and the speeds and voltages of everything, hell make it interactive
|
|
#
?
Apr 17, 2023 17:53
|
|
|
Perplx posted:my board has a com port header doing nothing, just send the whole boot log over that like linux does, you could get real detailed, like tell you what dimms were detected and the speeds and voltages of everything, hell make it interactive linux usually supports actual TTYs still. That's why virtual terminals are like TTY7 or whatever. I saw someone printing a neofetch recently (  ) it would actually be funni to see a real TTY running modern unix.
|
|
#
?
Apr 17, 2023 18:07
|
|
|
https://www.youtube.com/watch?v=dGbW7orZS-A Just a reminder that the silicon lottery still alive and well. In a test with 13 7600s, the difference between the best and worst performers in efficiency (measured as FPS/W) was 50%. Raw wattage numbers were 60.8w vs 85.7w, and the high watt part was also slower with the lowest clocks. Pretty huge difference.
|
|
#
?
Apr 17, 2023 18:39
|
|
|
Cygni posted:https://www.youtube.com/watch?v=dGbW7orZS-A that seems pretty extreme
|
|
#
?
Apr 17, 2023 18:43
|
|
|
makes sense. 7600 and 7900 is what’s left after binning for efficiency and taking everything that can be used and is needed for epyc, x3d, 7950, and 7700 (probably no difference with x). everything that fits requirements fits in the same way, everything that fails fails differently. you’re gonna have chiplets where 8 cores can hit identical bad clocks and at unacceptably high total heat and also ones where 2 cores are just dead but the other ones are golden. so most of the variation is going to be concentrated in that sku
|
|
#
?
Apr 17, 2023 20:21
|
|
|
Did the UK get any stock of 7800X3D's for launch at all? I've not heard a peep of anyone grabbing one and many retailers saying they've still to get their initial shipment.
|
|
#
?
Apr 17, 2023 20:22
|
|
|
Cygni posted:https://www.youtube.com/watch?v=dGbW7orZS-A Intel reserves their absolute best chips for the BlackCore systems they sell to high-frequency traders. EDIT: I didn't realize Exacta (the company Intel bought, and which builds BlackCore systems) started selling AMD systems too 
|
|
#
?
Apr 17, 2023 22:26
|
|
|
BlankSystemDaemon posted:Did anyone ever say it wasn't? I think if you asked a random hardware enthusiast whether they thought there was still a 50% difference in performance per watt between identically labeled chips, they would have said no.
|
|
#
?
Apr 17, 2023 22:54
|
|
|
Cygni posted:I think if you asked a random hardware enthusiast whether they thought there was still a 50% difference in performance per watt between identically labeled chips, they would have said no. 
|
|
#
?
Apr 17, 2023 23:16
|
|
|
Cygni posted:I think if you asked a random hardware enthusiast whether they thought there was still a 50% difference in performance per watt between identically labeled chips, they would have said no.
|
|
#
?
Apr 18, 2023 00:50
|
|
|
BlankSystemDaemon posted:Surely they know about binning? I'm a huge PC enthusiast and know about binning and a 50% variance across the same make/model is crazy to me. edit:  Imagine getting 50% variance in MPG between the same car make/model.
|
|
#
?
Apr 18, 2023 01:11
|
|
|
The video serves as a good reminder for veterans and a primer for newer PC builders who may not already know about the silicon lottery. No need to be all "We knew this already  " about it
|
|
#
?
Apr 18, 2023 01:14
|
|
|
I'm going to watch the video because yeah, I'm shocked by that much variance and I've been overclocking for ages. I wonder if PBO is blasting tons of power in there to get no further than a manual overclock could have.
|
|
#
?
Apr 18, 2023 01:18
|
|
|
50% is a hell of a lot. Was the gap always that big?
|
|
#
?
Apr 18, 2023 01:22
|
|
|
After skimming the video, the temperatures were not correlated with power going into the chip, which can only mean that there were differences in thermal transfer to the cooler. All the old tricks can help there, like lapping the heat spreader or delidding. We knew that delidding helped a ton: https://www.youtube.com/watch?v=y_jaS_FZcjI This is probably because of big variances in thermal paste inside of the extra thick AM5 spreader. Back in my day, the dies were soldered to the heat spreader so no, there was not this much variance.
|
|
#
?
Apr 18, 2023 01:32
|
|
|
Twerk from Home posted:Back in my day, the dies were soldered to the heat spreader so no, there was not this much variance. back in my day we cracked the die with a heatsink that we attached with a screwdriver that slipped off and scraped some caps off of the motherboard
|
|
#
?
Apr 18, 2023 01:38
|
|
|
In my day we lubed up an abacus
|
|
#
?
Apr 18, 2023 01:49
|
|
|
Inept posted:back in my day we cracked the die with a heatsink that we attached with a screwdriver that slipped off and scraped some caps off of the motherboard 
|
|
#
?
Apr 18, 2023 05:36
|
|
|
The actual working part of the silicon is on the bottom, so any knicks or scratches on the topside shouldn't completely brick the chip.
|
|
#
?
Apr 18, 2023 05:43
|
|
|
Definitely cracked a core so bad on an Athlon XP that it shorted and left an etch of the writing on the giant copper slug heatsink I had.
|
|
#
?
Apr 18, 2023 05:49
|
|
|
Would just like to express my annoyance at gigabyte's stock fan curves and fan control functionality in their bios. Who thought it would be a good idea to max all of your fans out whenever the CPU touches 60C? On every single preset, including the "silent" preset! They also don't let you copy fan curves between fans, so you have to manually input a custom curve for every fan.
|
|
#
?
Apr 18, 2023 06:54
|
|
|
Twerk from Home posted:After skimming the video, the temperatures were not correlated with power going into the chip, which can only mean that there were differences in thermal transfer to the cooler. All the old tricks can help there, like lapping the heat spreader or delidding. My thoughts are that regardless of which of those steps are in play though, the end user still apparently has a much bigger power and performance variation window to be aware of than I think most people thought. Thats something to remember.
|
|
#
?
Apr 18, 2023 08:31
|
|
|
the other thing that comes to mind in light of that video is that when product segmentation can come down an alleged spread of no more than 300 Mhz, how are manufacturers ensuring that there's actual value to these SKUs? presumably you wouldn't ever see a 7600 non-X crossing into 7600X territory because in this particular case the power targets and clock speed difference is so much higher, but were there people who were getting 3600X, or even 3600XT-level performance out of their 3600s? is there any kind of validation in play to ensure that a 3600XT is/was always distinctly faster than its two preceding variants?
|
|
#
?
Apr 18, 2023 08:43
|
|
|
Dr. Video Games 0031 posted:Would just like to express my annoyance at gigabyte's stock fan curves and fan control functionality in their bios. Who thought it would be a good idea to max all of your fans out whenever the CPU touches 60C? On every single preset, including the "silent" preset! They also don't let you copy fan curves between fans, so you have to manually input a custom curve for every fan. I don't know if they changed it on AM5, but on my X570 Aorus Elite BIOS I can set one set of curves and apply it to multiple different fans. It's not the most intuitive design, and if I remember rightly (I'm not at home right now) it is a pop-up window that lets you select multiple fans from the various options. I think you can apply the same curve to all your fans, except maybe the chipset fan (which I don't think the newer boards have, anyway).
|
|
#
?
Apr 18, 2023 08:50
|
|
|
gradenko_2000 posted:presumably you wouldn't ever see a 7600 non-X crossing into 7600X territory because in this particular case the power targets and clock speed difference is so much higher, but were there people who were getting 3600X, or even 3600XT-level performance out of their 3600s? is there any kind of validation in play to ensure that a 3600XT is/was always distinctly faster than its two preceding variants? At one point my PC had a 3600x in it and my wife’s had a 3600. The 3600 ran cooler and seemed to held boost clocks better than my 3600x after setting +200MHz. Small differences but it was there. There was very little difference between the 3600, X and XT out of the box so cooling and things like how much voltage the IOD/RAM needed could easily dominate.
|
|
#
?
Apr 18, 2023 13:40
|
|
|
When I set up this X570S Aorus Master I definitely had a "copy curve to all fans" button and I think I could pick which fans "all" meant from there, but it was also definitely the opposite of intuitive. Maybe it's just the X-series boards have the fun fan curve buttons?
|
|
#
?
Apr 18, 2023 13:40
|
|
|
I think both my Gigabyte B450 and B550 ITX boards have this, too.
|
|
#
?
Apr 18, 2023 13:53
|
|
|
|
| # ? May 15, 2024 13:19 |
|
|
Dr. Video Games 0031 posted:Would just like to express my annoyance at gigabyte's stock fan curves and fan control functionality in their bios. Who thought it would be a good idea to max all of your fans out whenever the CPU touches 60C? On every single preset, including the "silent" preset! They also don't let you copy fan curves between fans, so you have to manually input a custom curve for every fan.
|
|
#
?
Apr 18, 2023 14:05
|
|