Speculation: AMD's response to Intel's 8-core i9-9900K

Page 12 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

AtenRa

Lifer
Feb 2, 2009
14,003
3,361
136
Ironically this really only seemed to be a problem for AMD which has more PCIe lanes from the CPU.

Pretty much the opposite of the original claim surrounding Intels lack of PCIe lanes from the CPU.

Intels chipset based lanes actually ran just as fast as AMDs CPU lanes. It's only AMDs chipset lanes that slow down significantly.

X470 only has PCIe Gen 2, no wander it has half the speed than the CPU lanes. Nobody will use the chipset lanes when you can use the Gen 3.0 lanes from the CPU.
 

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
X470 only has PCIe Gen 2, no wander it has half the speed than the CPU lanes. Nobody will use the chipset lanes when you can use the Gen 3.0 lanes from the CPU.

That doesn't alter my point.

Someone brought up Intels Lower CPU PCIe lane count, and the supposed detrimental effect on it's storage access speed as a slam against Intel CPUs.

Turns out, the reality was essentially the opposite.
 
Reactions: Accord99

ub4ty

Senior member
Jun 21, 2017
749
898
96
That doesn't alter my point.

Someone brought up Intels Lower CPU PCIe lane count, and the supposed detrimental effect on it's storage access speed as a slam against Intel CPUs.

Turns out, the reality was essentially the opposite.
The devil's in the details which I've noticed you don't like probing.
They purposely only tested one NVME drive in their tests. Intel hangs its pcie 3.0 extensions beyond the first two slots pcie slots (x16 or x8/x8) off of their chipset. The chipset is DMI 3.0 (3.93GB/s) .. PCIE 3.0 x4 is (3.94GB/s). Yes, the whole chipset is serviced by a PCIE 3.0 x4 link. The problem/slight of hand is obvious. By only testing one NVME drive, they keep the shared channel from being saturated. This is why AMD hung the NVME off of a dedicated x4 slot directly connected to the CPU. Only the 2nd one is serviced through a much slower link through the chipset. My understanding there is that the interface is gimped down to PCIE 2.0 for AMD's X370 to ensure it doesn't slam the chipset's connection when active. The test gave the expected results. The detrimental effects in storage access speed on Intel come when you begin using other chipset features (something the reviewers here purposely avoided) and didn't detail. This doesn't happen using the proper M.2 slot on AMD's board because its directly connected to the CPU.

INTEL Z370 CHIPSET Connectivity :


Intel went full retard on oversubscription through the chipset.
AMD decided to recognize NVME as a first class citizen and give it a dedicated x4 connection.
In AMD's case, through the chipset, you have a connection to a second M.2 but you're going to pay a performance penalty so you don't saturate the chipset link where tons of other peripherals hang off. The consumer can decide which model works best for them but the linked review is a joke and flat out omits these details... Such is the case for information curated for the masses.
 

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
The devil's in the details which I've noticed you don't like probing.
They purposely only tested one NVME drive in their tests. Intel hangs its pcie 3.0 extensions beyond the first two slots pcie slots (x16 or x8/x8) off of their chipset. The chipset is DMI 3.0 (3.93GB/s) .. PCIE 3.0 x4 is (3.94GB/s). Yes, the whole chipset is serviced by a PCIE 3.0 x4 link. The problem/slight of hand is obvious. By only testing one NVME drive, they keep the shared channel from being saturated. This is why AMD hung the NVME off of a dedicated x4 slot directly connected to the CPU. Only the 2nd one is serviced through a much slower link through the chipset. My understanding there is that the interface is gimped down to PCIE 2.0 for AMD's X370 to ensure it doesn't slam the chipset's connection when active. The test gave the expected results. The detrimental effects in storage access speed on Intel come when you begin using other chipset features (something the reviewers here purposely avoided) and didn't detail. This doesn't happen using the proper M.2 slot on AMD's board because its directly connected to the CPU.

INTEL Z370 CHIPSET Connectivity :


Intel went full retard on oversubscription through the chipset.
AMD decided to recognize NVME as a first class citizen and give it a dedicated x4 connection.
In AMD's case, through the chipset, you have a connection to a second M.2 but you're going to pay a performance penalty so you don't saturate the chipset link where tons of other peripherals hang off. The consumer can decide which model works best for them but the linked review is a joke and flat out omits these details... Such is the case for information curated for the masses.

It's funny that whenever it comes to slamming Intel over some supposed deficit, that depends more on hearsay than evidence, the same names come up over and over again.

Instead of evidence, we just get your claims that "Intel went full retard", which not what I consider in any way convincing.

Instead of long winded anti-Intel diatribes. How about some evidence.
 
Reactions: Zucker2k

Abwx

Lifer
Apr 2, 2011
11,167
3,862
136
It's funny that whenever it comes to slamming Intel over some supposed deficit, that depends more on hearsay than evidence, the same names come up over and over again.

Instead of evidence, we just get your claims that "Intel went full retard", which not what I consider in any way convincing.

Instead of long winded anti-Intel diatribes. How about some evidence.

He gave your full of evidence, how about answering his technical claims rather than posting an erratic and unrelated post..?.
 

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
He gave your full of evidence, how about answering his technical claims rather than posting an erratic and unrelated post..?.

He gave ZERO evidence. He gave a long diatribe.

Evidence would be links to the benchmarks demonstrating his claims.
 

ub4ty

Senior member
Jun 21, 2017
749
898
96
It's funny that whenever it comes to slamming Intel over some supposed deficit, that depends more on hearsay than evidence, the same names come up over and over again.

Instead of evidence, we just get your claims that "Intel went full retard", which not what I consider in any way convincing.

Instead of long winded anti-Intel diatribes. How about some evidence.
> Provides ample evidence and a diagram
> Peter Scott ignores it and does what he claims I'm doing
Wew lad...
I bet it's no worse than what happens to the NVMe drive in the second slot on AMD X470:
https://www.techspot.com/review/1646-storage-performance-intel-z370-vs-amd-x470/

PCIE 3.0 x4 + PCIE 2.0 Bandwidth (AMD NVME slot #1 + Chipset slot #2) > PCIE 3.0x4 bandwidth through DMI 3.0
This is just basic math. Extended analysis reveals that NVME versioning keeps improving as does its throughput. Eventually, you will be able to achieve full saturation. So, do you provision for this like AMD does and give a first class high speed storage its own dedicated connection to the CPU or do you cobble everything together in the system through one chipset straw and cause every device's performance hanging off of it to degrade because your first class citizen NVME is hogging the bandwidth of a shared channel? I'm fine with the split AMD made. It is the most sensible. My main drive is NVME and I want it performing at max capacity w/o hinderance. The rest is secondary storage. I don't have M.2 slot #2 occupied because that's what SSD is for. I also have other peripherals on the chipset that need the bandwidth it offers. I want it sensibly shared.

Intel isn't a catch up company. They were the leader in this space and even have pushed the envelope with Optane storage. Yet, they keep up w/ this idiotic chipset affair and PCIE lane count gimping.

This continues all the way up into their X299 HEDT platform :
There's been some confusion regarding the amount of lanes on x299 vs x399. For the record:
For x399, AMD has provided a total of 64 lanes - 60 from the CPU, 4 from the chipset.

http://www.anandtech.com/show/11482/amd-cpu-updates-threadripper-64-pcie-lanes-epyc-june-20th
x299 has a total of 68 lanes for the 10c+ CPUs: 44 from the CPU, 24 from the Chipset
Intel is connecting the CPU via DMI to the chipset. This has PCH lanes, which are equivalent to a PCIe3 x4. This means all 24 PCIe lanes from the chipset have to share the PCH (PCIe3 x4) link to the CPU.

That 24 is squeezed down to 4 because that's all its wired for.
So, x299 has 48 lanes.

The ride never ends with intel
. This is the problem.
Meanwhile, I actually do have 3 NVME drives on my 399 board and its handling very seriously read/writes. So, this would be a serious problem on X299.
AMD : x4/x4/x4 - 3 nvme drives
Intel : Even though our platform costs more, lets go ahead and cram that all through one x4 pipe and for kicks lets put every other peripheral on the same pipe too

What the fudge?
 
Last edited:

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
> Provides ample evidence and a diagram
> Peter Scott ignores it and does what he claims I'm doing
Wew lad...

Uhh. No. You made claims.

I linked the benchmarks showing a direct relevant comparison. IOW Evidence.

This continues all the way up into their X299 HEDT platform :

Ah, classic ub4ty. Theory crafting and goalpost shifting. Never evidence.
 

ub4ty

Senior member
Jun 21, 2017
749
898
96
Uhh. No. You made claims.
I linked the benchmarks showing a direct relevant comparison. IOW Evidence.
Ah, classic ub4ty. Theory crafting and goalpost shifting. Never evidence.
I would like to take you out to dinner one evening Mr.Scott.
The steak and wine will be on me and rendered by an AMD based compute stack.
 
Reactions: Markfw

AtenRa

Lifer
Feb 2, 2009
14,003
3,361
136
That doesn't alter my point.

Someone brought up Intels Lower CPU PCIe lane count, and the supposed detrimental effect on it's storage access speed as a slam against Intel CPUs.

Turns out, the reality was essentially the opposite.

The reality that you dont understand is that if you use 2x drives at the same time,

In Intels Z370/390 the speed will be half on both drives vs a single drive.
In AMDs X370/470 one drive will continue to have full speed and only the second drive will have half the speed.

Got it now ???
 

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
The reality that you dont understand is that if you use 2x drives at the same time,

In Intels Z370/390 the speed will be half on both drives vs a single drive.
In AMDs X370/470 one drive will continue to have full speed and only the second drive will have half the speed.

Got it now ???

The desktop PC use case for saturating more than one fast NVMe drive?
 

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
It's funny how the general feeling of those on the Intel platform is actually positive against the negative stance of those who won't touch an Intel chip with a 10-foot pole. Interesting.
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,361
136
The desktop PC use case for saturating more than one fast NVMe drive?

Many, you have one M2.0 drive and then you have a Secondary large SATA/USB 3.0/1 drive for storage. The moment you are going to transfer/copy from one to the other or to a third drive, your primary M2.0 on Intel Z370/390 will have half the speed.

And as prices going down, you may start with a single M2.0 drive but get a second or even third down the road. Then your Intel Z370/390 platform starts to have a serious bottleneck.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,740
14,772
136
It's funny how the general feeling of those on the Intel platform is actually positive against the negative stance of those who won't touch an Intel chip with a 10-foot pole. Interesting.
And what of those of us who are on BOTH platforms ? I have 9 Intel systems, and 7 AMD systems, and core counts wise, more Intel cores total.

Reality and truth are sometimes hard to accept for some.
 

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
And what of those of us who are on BOTH platforms ? I have 9 Intel systems, and 7 AMD systems, and core counts wise, more Intel cores total.

Reality and truth are sometimes hard to accept for some.
How old are your Intel systems, and has the manufacturer released bios updates against spectre, etc?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,740
14,772
136
How old are your Intel systems, and has the manufacturer released bios updates against spectre, etc?
Most are socket 2011-3, Xeon E5-2683v3. I also have a 2600k, 4790k, 6700k, Xeon 5570 and a socket 775 Xeon thats not turned on anymore.

Since most of them are all linux and just doing DC work, I have not pursued any updates.

Edit: And I should say, my 2 2700x systems (8c/16t) at stock are beating my E5-2683v3 Xeons (14c/28t) at points in DC (Rosetta/WCG) at less power usage.
 
Last edited:

Abwx

Lifer
Apr 2, 2011
11,167
3,862
136
It's funny how the general feeling of those on the Intel platform is actually positive against the negative stance of those who won't touch an Intel chip with a 10-foot pole. Interesting.

That s quite a total inversion of reality, as you ll hardly find an AMD user who never owned an Intel CPU, not sure that the reverse hold as well, or does it ?..
 
Reactions: ub4ty

ub4ty

Senior member
Jun 21, 2017
749
898
96
And what of those of us who are on BOTH platforms ? I have 9 Intel systems, and 7 AMD systems, and core counts wise, more Intel cores total.

Reality and truth are sometimes hard to accept for some.
Same here. I have just as many intel rigs as I have Ryzen rigs. Albeit, the core counts are different.I rode Intel all the way up to quad core/8thread. AMD won on 8core+ so I switched over. What one would expect for someone who has zero bias. I switch teams whenever there is better value as any mindful consumer should. I gave AMD's new RX Vega video card a shot but the driver/dev stack was shot for my purposes, so I sold it like a hot cake and went straight back to Nvidia. The pure intel bots don't understand this.

Many, you have one M2.0 drive and then you have a Secondary large SATA/USB 3.0/1 drive for storage. The moment you are going to transfer/copy from one to the other or to a third drive, your primary M2.0 on Intel Z370/390 will have half the speed.

And as prices going down, you may start with a single M2.0 drive but get a second or even third down the road. Then your Intel Z370/390 platform starts to have a serious bottleneck.
Also, a big trend that has moved down to the desktop space is tiered storage where you have a caching policy across RAM/NVME/SSD/HDD. AMD released software named FuzeDrive. Windows Server has it built in and there are tons of other software packages. Meanwhile, this was the whole thinking behind Intel's Optane drives. You can easily begin to saturate PCIE 3.0 x4<->RAM with a cache policy.
 
Reactions: Drazick and Markfw

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,740
14,772
136
Same here. I have just as many intel rigs as I have Ryzen rigs. Albeit, the core counts are different.I rode Intel all the way up to quad core/8thread. AMD won on 8core+ so I switched over. What one would expect for someone who has zero bias. I switch teams whenever there is better value as any mindful consumer should. I gave AMD's new RX Vega video card a shot but the driver/dev stack was shot for my purposes, so I sold it like a hot cake and went straight back to Nvidia. The pure intel bots don't understand this.
Thats where all my E5-2683v3 Xeons came from, until Ryzen hit. They were $350 on ebay and great performers. But now my 8 core Ryzens are stomping their 14 cores, so Ryzen it is. Oh, and with less power draw.
 
Reactions: Drazick

ub4ty

Senior member
Jun 21, 2017
749
898
96
Thats where all my E5-2683v3 Xeons came from, until Ryzen hit. They were $350 on ebay and great performers. But now my 8 core Ryzens are stomping their 14 cores, so Ryzen it is. Oh, and with less power draw.
Yep, I'm a big fan of buying enterprise hand me down hardware when they do big upgrades.I was all over the enterprise Xeon build for my upcoming build and then Ryzen came out. Then AMD went scorched earth and released Thread ripper... My God, it's like the heavens parted when Thread ripper was released. It was exactly what I had been dreaming of and needed for my upcoming build. The far lower power draw and the insane affordability were a godsend. It changed the face of computing IMO. I recall Nvidia even thanking AMD for changing the game with regards to far more PCIE slots. A whole new range of possibilities exists because of the Ryzen architecture.

Intel is no slouch and I fully expect their integration of FPGA, Optane, EMIB to be a game changer. They simply need to get real about pricing and behave like there fighting for people's dollars. Until them, I'm panning them all together. They milked me enough over the years.
 
Reactions: lightmanek

PeterScott

Platinum Member
Jul 7, 2017
2,605
1,540
136
It's funny how the general feeling of those on the Intel platform is actually positive against the negative stance of those who won't touch an Intel chip with a 10-foot pole. Interesting.

The fact of Intel catching up on core count really brings out the AMD partisans to denigrate Intels part before it arrives.
 
Reactions: whm1974
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |