Contribute
Register

Z490 & Z590

Joined
Feb 16, 2012
Messages
180
Motherboard
Z590i Vision D
CPU
i7-11700
Graphics
RX560
Mac
  1. MacBook Air
Wow thanks for the totally unneeded graph. I understand PCI-E x16 is far faster then TB3 if it wasn't TB3 PCI-E add-on card would not function in a x4 slot. And the Mac Pro 6,1 was not a system designed with the ability to added PCI-E devices so logically it was not crippled. Designed inferior to the Mac Pro 5,1 yes but not crippled.
It's not true. PCIe has been replaced by Thunderbolt 2 in Mac Pro 6.1. Unfortunately, it was not a good idea.
Besides, few people realize that Thunderbolt 3 has only theoretically 40Gbit bandwidth. In fact, these 40Gbit are divided into DATA and Video. Which gives us max. 22Gbit to DATA. Which is less than PCIe x4.
transfery.png
 
Joined
Jul 13, 2016
Messages
2,818
Motherboard
Gigabyte Z370 Gaming 5
CPU
I9 9900k
Graphics
Vega 64
Mac
  1. Mac mini
  2. Mac Pro
Mobile Phone
  1. iOS
It's not true. PCIe has been replaced by Thunderbolt 2 in Mac Pro 6.1. Unfortunately, it was not a good idea.
Besides, few people realize that Thunderbolt 3 has only theoretically 40Gbit bandwidth. In fact, these 40Gbit are divided into DATA and Video. Which gives us max. 22Gbit to DATA. Which is less than PCIe x4.
Not even sure what you are talking about you are totally off topic of this thread take your needless graphs and charts that are totally not relevant or needed in this tread somewhere else like maybe a dark corner.
 
Joined
Feb 16, 2012
Messages
180
Motherboard
Z590i Vision D
CPU
i7-11700
Graphics
RX560
Mac
  1. MacBook Air
Not even sure what you are talking about you are totally off topic of this thread take your needless graphs and charts that are totally not relevant or needed in this tread somewhere else like maybe a dark corner.
You're wrong. Bringing up the topic of transfers is absolutely necessary in this thread. I took the last photo from Intel's detailed specification:
The actual, realistically possible transfers after Thunderbolt 3 are shown there.
People have been misled into thinking Thunderbolt = PCIe. And that's not true. Thunderbolt is good as an extension of the possibilities, but not as a replacement.
Therefore, building a hackintosh still makes sense, because Apple has stumbled into a dead end of miniaturization.

PC MINI vs Mac MINI.png
 
Joined
Jul 13, 2016
Messages
2,818
Motherboard
Gigabyte Z370 Gaming 5
CPU
I9 9900k
Graphics
Vega 64
Mac
  1. Mac mini
  2. Mac Pro
Mobile Phone
  1. iOS
You're wrong. Bringing up the topic of transfers is absolutely necessary in this thread. I took the last photo from Intel's detailed specification:
The actual, realistically possible transfers after Thunderbolt 3 are shown there.
People have been misled into thinking Thunderbolt = PCIe. And that's not true. Thunderbolt is good as an extension of the possibilities, but not as a replacement.
Therefore, building a hackintosh still makes sense, because Apple has stumbled into a dead end of miniaturization.

View attachment 522463
You have to be the biggest idiot there is this thread is about z490 and z590 hacks not real apple miniature systems or thunderbolt. You are trying to jump the track with your own agenda. Take your thunderbolt discussion to its own thread.
 
Joined
Apr 12, 2021
Messages
63
Motherboard
Asus ROG z590 Maximus XIII Hero
CPU
i9-10900K
Graphics
W5700 Pro
Mac
  1. MacBook Pro
  2. Mac mini
  3. Mac Pro
Classic Mac
  1. Centris
  2. Power Mac
Mobile Phone
  1. iOS
That PC vs Mac mini graphic says it all Lol

But the question is begged by that same graphic as to what all that sweet PC mini x16 GPU mojo is actually good for... It's not like Appe's product designs don't account for the GPU: they do! Apple is not a meaningful part of the "PC gamer" market, and they serve the content and engineering markets with "workstation" GPUs that are tuned, via generally effective vertical integration, to applications that are valuable to those markets, so much so that for a good while, Apple owned the "desktop publishing" and NLE vid markets and made a good stab at taking over Hollywood.

The point is you just don't buy a Mini if you need GPU horsepower, and a PC Mini (as shown) is not miniature! So the whole thing goes in circles.

The points about TB limitations are fine, and welcome and correct info. Making the system bus wires long has a cost and TB is an interesting design compromise. The layman has to face that a eGPU in an external enclosure is crippled by TB compared to the GPU module design intentions, and this gets more clear in Apple's own SoC integrated RAM designs, which work well precisely because they obviate the PCI bus altogether! Nes pas?

As to flame wars on web: Flame on! Always enjoyable... Except when they're not.
 
Joined
Feb 16, 2012
Messages
180
Motherboard
Z590i Vision D
CPU
i7-11700
Graphics
RX560
Mac
  1. MacBook Air
That PC vs Mac mini graphic says it all Lol

...
The PC is smaller and lighter, despite the fact that it contains the same components inside, which are attached to the MINI mac on the right side. The mess around mac is the most terrifying thing. :p
 
Joined
May 11, 2011
Messages
1,612
Motherboard
Gigabyte B550 Vision D
CPU
Ryzen 5900X
Graphics
RX 6800
Mac
  1. iMac
  2. MacBook
  3. MacBook Air
  4. MacBook Pro
  5. Mac mini
  6. Mac Pro
Classic Mac
  1. Power Mac
  2. PowerBook
Mobile Phone
  1. iOS
A few things I'd thought to mention.

It's come to my attention that we may be finally getting some sort of a working fix for Ice Lake and potentially Tiger Lake G7 GPUs. Someone has posted a working build of the Iris G7 IGPU on Ice Lake Catalina, which uses the new OC Apple GUC firmware fix to work. > https://github.com/leon0410898/XPS13-9300-hackintosh


Screen Shot 2021-06-23 at 11.46.41 AM.png


On a separate note, Intel recently posted their latest June 2021 specification update about some of the issues affecting Rocket Lake chips, and it looks pretty serious. Some of the issues are PCIe and sleep related, others are initialising cache issues, incorrect L2 cache reporting and warm reset memory issues. The biggest seems to be the failing of the jitter tolerance compliance test which may affect Thunderbolt sustained bandwidth performance.

Screen Shot 2021-06-23 at 12.30.20 PM.png


Screen Shot 2021-06-23 at 12.00.24 PM.png

Screen Shot 2021-06-23 at 12.00.43 PM.png

Screen Shot 2021-06-23 at 12.23.51 PM.png
 
Joined
Apr 5, 2016
Messages
946
Motherboard
GIGABYTE Z170X-Gaming 7
CPU
i7-6700K
Graphics
GTX TITAN X
Mac
  1. iMac
  2. MacBook
  3. MacBook Air
  4. MacBook Pro
  5. Mac mini
  6. Mac Pro
Classic Mac
  1. Apple
  2. LC
  3. Power Mac
Mobile Phone
  1. iOS
PCIe link width transition is interesting. There's a way to programatically change the PCIe link rate (so you can test a PCIe card at different link rates). Is there a way to programatically change the PCIe link width?
 
Joined
Apr 12, 2021
Messages
63
Motherboard
Asus ROG z590 Maximus XIII Hero
CPU
i9-10900K
Graphics
W5700 Pro
Mac
  1. MacBook Pro
  2. Mac mini
  3. Mac Pro
Classic Mac
  1. Centris
  2. Power Mac
Mobile Phone
  1. iOS
PCIe Throughput by Gen.png


When TB was introduced, its performance potentials seemed pretty great, like omg (fantasy) PCI bus moves outside the box with bold new config possibilities for heterogeneous device clusters.

Apple seemed to me to implicitly appeal to such a fantasy by consolidating their small-system IO into a common port, making their designs seem svelte on one hand and expandable on the other. Simple and flexible. Seeing thr display connection unified with the USB connection egged on (unconsciously to me) my fantasy because the most important and demanding consumer PCIe device is the display adapter and seeing the hybrid emerge seemed cool.

But the payoff ultimately seems blasée. Look at the chart above, and surmise TB evolution with PCIe evolution. Duh. By simple traits of physics, a long-wire external bus is never gonna keep up with its internal short-wire parallel parent design.

But the chart also shows something that maybe should have been predicted, but is also an unexpected development: PCIe gen 3 stagnated on Intel for 11 years whereas the projected pace of evolution was more like 3.5 per generation.

The mess which is TB4 today is still about PCIe x4 for most users due to this stagnation.

OTOH PCIe gen 3 x4 I/O seems like nothing to sneeze at, a good middle ground for consumer fare.

Yet the projection was TB would be PCIe x1 by this date? Why did it stagnate? The designers must know all this.

PC gaming gfx was not a harbinger of UI evolution. Putting xparency and zoom fx in UI is not challenging compared to 3d first person open world gaming. The killer app for consumer gfx turned out to be video decode for increasing display resolutions. This was developed in concert with the big entertainment companies and doesn't feel like a killer app to consumers 'cause generations see TV as a god-given right and the PC replaced the TV — literally, now every TV has a PC-thing inside of it. This is very important evolution, but also boring b/c its a pure engineering matter which can be solved and budgeted, per Intel iGPU.

So Thunderbolt sort of looked to me like it was breaking out I/O for about 5 years and this seemed vaguely promising.

Yet architecturally it's lead to almost nothing of interest. USB keeps up with it in practice. NVMe storage can use its throughput, but somehow this doesn't translate into any capability that feels like a break-through for peripheral storage; it's not enabling anything that wouldn't otherwise be possible except for shuffling hi-res web video to an external drive from minimally provisioned base-system drives. And this brings me back to TV again (what does web vid mean for most of us? incessant contact with cam-whores of infinite stripes).

As aside, NVMe while being solidly 5x faster than SATA does not make any PC feel 5x faster, it just keeps the UI perky enough to feel like its not standing in your way. Meanehile, UIs feel completely insane since Web 2.0 but like crappy TV shoes, everyone enjoys the novelty.

Gaming is less interesting for tech than its fan-base would make you think — or as just more cam-whores does make me think. It has been perfected to the point that streaming is eclipsing dedicated gaming rig, except for the Xtremesports set. GPUs are in short supply, but not because of UI or content creation or gaming revolution, but to feed vast Ponzi scheme of crypto-asset production.

Moreover, we are in this paradoxical passage of history where PC HW has somehow advanced even beyond the capacity of SW bloat, yet we still feel we don't have enough. Is there any incredible new tech less compelling than 8K displays?!

Bsck to TB: the long wires aren't that long, and they cost a fortune. Today it costs an additional $300-400 US to connect a pair of displays and a backup drive to a $1200 macbook via a Caldigit TB4 hub, and the Macbook doesn't support more than one ext display. Cables are insanely expensive. You can buy a fully provisioned NUC with Windows for same $$ as TB expansion

At same time a dGPU can be connected to a mini via TB but the iGPU is engineered to run most apps.
High-end vid producers will buy dedicated workststions a la MacPro. Gamers are on PC.

As a modern hackintosher I am keenly aware failed promise of mature TB and the tedium of even thinking about it, because it obviously does almost nothing that isn't covered by pre-existing tech. (special props to Audio nerds who are a special herd of cats who pray at an alter of sensory rapture contrary to everything that vid production gurus have to teach about system configuration.) We already have USB and display ports. If you have a desktop, you have PCIe for nerding out. And if you have a non-Apple portable... Wait, why do you have a non-Apple portable but want to run MacOS? By nature you are a haploid whose PC genetic inheritance has incurred upon you a perverse need for pain.

So TB amounts to a very expensive display port cable. And the promise of TB is delivered as a fashionably minimal connector arrangements on Apple devices sold to a market of sexy vloggers. The end result being the cultural triumph of the anonymous youtube front page and Qanon.

At one time a long time ago there was hope that TB-like interconnect would commodify explicitly parallel heterogenious multicomputer architecture to enable very flexible and powerful systems to be scaled out of inexpensive modules. Curiously, it was discovered that no-one really cares about such a holy-grail of computing. The future has always preferred just packing more logic into the chips.

It now completely clear TB is Apple-specific value-add, and with TB4 it looks very close to being an Apple-branded feature where there is reason to be concerned that TB peripherals will all be vendor specific: Is that an Apple TB or a Dell TB? It's not a super big deal; this has always been true. But when Apple was making its kit from PC parts there was a crossover between worlds which created consumer possibilities. Now its gonna be Apple and MSFT fighting over the entire stack again top to bottom, like the dayz when that Pepsi idiot took over Apple.

Does either company have a vision for the future beyond stuffing their markets in what will be doomed to be the caskets called their "stores"?

Linux?

I've never had less hope in the promise of personal computing.
 
Joined
Jun 28, 2021
Messages
6
Motherboard
Gigabyte Z590 Vision D
CPU
Intel i9 10900
Graphics
Radeon VII
Has anyone had a strange issue with on-board audio (Z590 Vision D), where the balance and volume get changed on a cold boot? I'm guessing my usbmap must be wrong, but I didn't have anything to test the internal headers with.
 
Top