16 Core 32 Thread HP z820 dual socket workstation build/performance upgrade

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
LOL, that's because CPU-Z which you are using for benchmarking is unable to make more than 70W load on 2650v2. And I'm talking about Linx that pulls about 110W from 2650v2 (at 3GHz!) during first 10 seconds, then 'Power Limit 1' comes into play and CPU is not permitted to consume more than TDP hence dropping its frequency below 2.9GHz.
Oh Boys! Linux comes into play, another can of worms lol. FYI I was able to easily hit 2673 v2 at 110 watts on the CPUz benchmark including monitoring with both hardware monitor and Core Temp. Is this a hardware limitation specific to the 2650v2, because from I gather, because all my other 2600 v2 chips reach their full TDP when being torture tested by CPUz. In other words, CPUz was hitting 95 watts with my 2650 v2 runs, and for some reason yours was not.
 
Last edited:

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
LOL, that's because CPU-Z which you are using for benchmarking is unable to make more than 70W load
Wait are you saying CPU-z is capping load during the torture test at 70 watts? I find this is very difficult to believe. Heck I just tested with like 4 different computers, including my 9600KF which topped out at the specified 98.4W... Please explain why the cpuz limits the xeon 2650 v2, because this is the first I've heard of this issue.
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
Oh Boys! Linux comes into play, another can of worms lol. FYI I was able to easily hit 2673 v2 at 110 watts on the CPUz benchmark including monitoring with both hardware monitor and Core Temp
LOL, not Linux, but Linx. Is it so difficult for you to download it, launch it and see how your 110w TDP CPU will barely make 3.1 GHz under real AVX load?:D And it will throttle, because if CPU consumes 110w under CPU-Z bench, it will require 140 under any AVX load (Linx for instance). And this is where 2687w v2 will do and 2673v2 will suck.

Wait are you saying CPU-z is capping load during the torture test at 70 watts?
No, this is simply not power-hungry load type for a CPU. High TDP versions of server CPUs were specially brewed to maintain multipliers under real power hungry computational AVX loads, while some OEM versions were designed to work under light load (Amazon, Facebook servers) but with high multipliers.
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
No, this is simply not power-hungry load type for a CPU. High TDP versions of server CPUs were specially brewed to maintain multipliers under real power hungry computational AVX loads, while some OEM versions were designed to work under light load (Amazon, Facebook servers) but with high multipliers.
This is complete speculation on your part, considering scores for the "lite" OEM CPUs are in some cases BETTER than the retail equivalent. LOL

The 2650 v2 is a 95 watt CPU, right? I was pulling 95 watts with CPU-z, so your testing is not working out right.

And what do you make of the geekbench scores? 2673 v2 actually comes out on top there.
*crickets* LOL

LOL, not Linux, but Linx. Is it so difficult for you to download it, launch it and see how your 110w TDP CPU will barely make 3.1 GHz under real AVX load?:D And it will throttle, because if CPU consumes 110w under CPU-Z bench, it will require 140 under any AVX load (Linx for instance). And this is where 2687w v2 will do and 2673v2 will suck.
Ever heard of something called AVX offset? LOL
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
There is a lot of mis-information out there on the internet boys. Be sure to do your research when purchasing your next CPU! For example, after doing some extensive research I determined the 2696v2 is actually a better processor than the flagship 2797 v2, intel got them mixed up for flagship CPU, LOL
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
LOL, that's because CPU-Z which you are using for benchmarking is unable to make more than 70W load on 2650v2.
LOL you do realize your logic here, right? You are essentially saying that the CPU-z benchmark / torture test cant peg all CPU cores at 100%. That's just ridiculous thinking and I've confirmed you are completely on a number of xeon rigs ad my disposal.
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
This is complete speculation on your part, considering scores for the "lite" OEM CPUs are in some cases BETTER than the retail equivalent
Yep, under some 'lite' type of load those OEM versions are better, but they sucks in heavy computational job. That is why they were never meant for public release. Generic customers asking: "why my Xeon sucks under load as much as common Celeron?" is not what Intel wishes most
The 2650 v2 is a 95 watt CPU, right? I was pulling 95 watts with CPU-z, so your testing is not working out right.
This time you are right. That is because I don't use Hyper Threading, most useless performance-wise feature, but, as you correctly noticed, +25W to power consumption.
Ever heard of something called AVX offset? LOL
Sure. AVX offset was introduced in Skylake, generations after your 2673v2 first saw the world. Your CPU cannot set different multipliers for different types of load, instead it can only throttle being limited by the TDP, while 2687w regardless of generation will never change frequency under AVX load
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
You are essentially saying that the CPU-z benchmark / torture test cant peg all CPU cores at 100%. That's just ridiculous thinking and I've confirmed you are completely on a number of xeon rigs ad my disposal.
What ridiculous is that you don't see the difference between 'peg all CPU cores at 100%' and power consumption. And that explains everything
 

MBastian

Active Member
Jul 17, 2016
205
59
28
Düsseldorf, Germany
What ridiculous is that you don't see the difference between 'peg all CPU cores at 100%' and power consumption. And that explains everything
I've put this guy on ignore a long time ago, unfortunately I am not rarely logged in and I just can't turn my eyes away from his drivel. He pops in every few month and fishes for replies and disputes. Please, don't feed the Troll.
 
Last edited:

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
LOL, that's because CPU-Z which you are using for benchmarking is unable to make more than 70W load on 2650v2. And I'm talking about Linx that pulls about 110W from 2650v2 (at 3GHz!) during first 10 seconds, then 'Power Limit 1' comes into play and CPU is not permitted to consume more than TDP hence dropping its frequency below 2.9GHz.
I have to admit you are "preporting" to be a tech guy so above all should know first hand that turbo operation is NOT limited to TDP on Ivy Bridge, Ever heard of PL2? It's the PL2 specification that determines that wattage the processor can use at all core turbo.

Yep, under some 'lite' type of load those OEM versions are better, but they sucks in heavy computational job. That is why they were never meant for public release. Generic customers asking: "why my Xeon sucks under load as much as common Celeron?" is not what Intel wishes most
This time you are right. That is because I don't use Hyper Threading, most useless performance-wise feature, but, as you correctly noticed, +25W to power consumption.

Sure. AVX offset was introduced in Skylake, generations after your 2673v2 first saw the world. Your CPU cannot set different multipliers for different types of load, instead it can only throttle being limited by the TDP, while 2687w regardless of generation will never change frequency under AVX load

I don't know how many times i am going to have to explain this. The fact is that Ivy Bridge is not limited by TDP. It is limited by PLL and thermal parameters. In that case
What ridiculous is that you don't see the difference between 'peg all CPU cores at 100%' and power consumption. And that explains everything
When you read power consumption from HWmonitor, Core Temp and AIDA64 that all show the processor is at 95 watts, your statement about cpu-z not being able to hold more than 70 watts does not have much credibility. That's all I am saying.

Lets try to keep the conversational respective, I enjoy these discussions / debate and hopefully we can learn together without attacking each other. I have more data coming I'd like you to take a look at.
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
Just bare with me. I have some more screenshots, particularly PL2 values I'd like you to take a look at. If I remember correctly, the PL2 value for the 2673 v2 is 137 watts (unlimited). This means, this is the real "TDP" threshold you have to cross before the processor will drop down to base speed. I can confirm this 100% with the research I've done. I just have to go out and get a coffee now, I'll be back later to follow up.
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
I've put this guy on ignore a long time ago, unfortunately I am not rarely logged in and I just can't turn my eyes away from his drivel. He pops in every few month and fishes for replies and disputes. Please, don't feed the Troll.
I don't fish for anything. There are just people spreading mis-information about these processors and I am correcting the erroneous data. It's nothing personal. In fact, it's a benefit for those who are potentially considering upgrading to one of these chips.
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
Yep, under some 'lite' type of load those OEM versions are better, but they sucks in heavy computational job. That is why they were never meant for public release. Generic customers asking: "why my Xeon sucks under load as much as common Celeron?" is not what Intel wishes most
This makes absolutely no sense at all. Benchmarks prove the OEM 2696v2 performs nearly identically to the retail 2797 v2. Not only that, the 2696v2 actually outperforms the flagship 2697 v2 in multi-core performance do to 100MHz faster all core turbo speed. Not only that, I've yet to hear you respond to the geekbench findings in which the 2673 v2 actually beat the 2687w in multi core performance, albeit a small margin, but still with your logic we should see like half that, right?
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
What ridiculous is that you don't see the difference between 'peg all CPU cores at 100%' and power consumption. And that explains everything
What does this even mean? Once again you are failing to take into account the fact that the 2673 v2 has an all core turbo of 3.6GHz (as does the 130w 2667 v2 and the fire alarm 150 watt 2687w v2). And the Power limit 2 on the 2673 v2 is rated at 137.5W with an unlimited time component. Meaning that it will not throttle back to "base clock" under any work load. So effectively, all three processors have the SAME base speed of 3.6GHz, given you have proper cooling in place. This negates any 100MHz advantage in base clock that the 3687w may have. Looks better on paper, but just the opposite in reality.

And here you have been telling everyone under the sun that the 2673 v2 turbo performance is limited by the 110 watt TDP. This snip should speak for itself, enough said:
1630764438574.png


what is ridiculous is that you think OEM CPUs are suited for "lite duty" but that they suck under heavy computational jobs. This is also a complete fallacy. The OEM 2696 v2 actually performs better than Intel's flagship 2697 v2 under heavy work loads due to 100MHz higher all core turbo.

Yep, under some 'lite' type of load those OEM versions are better, but they sucks in heavy computational job. That is why they were never meant for public release. Generic customers asking: "why my Xeon sucks under load as much as common Celeron?" is not what Intel wishes most
Im not even going to humor a response to that first sentence. And that's not why they were never meant for public release. OEM CPUs were intended for use in a wide range of hardware, a very diverse system environment since they are usually sold with servers that need to perform in a wide variety of operating conditions. Meaning they need to be more versatile than retail chips which typically fit a certain niche market. This is why OEM chips must strike a better balance between core count, frequency and voltage. And if you study the OEM processors in the 2600 series family you will see that I am not bluffing here.
 

Attachments

Whaaat

Active Member
Jan 31, 2020
301
157
43
If I remember correctly, the PL2 value for the 2673 v2 is 137 watts (unlimited).
LOL. PL2=forever is a feature of motherboard vendors that doesn't grant Intel recommendations, not a particular CPU feature. And if your motherboard set tau=unlimited value automatically or allows you to set it manually doesn't mean this CPU will behave similarly in other motherboards such as Supermicro or Dell.
This means, this is the real "TDP" threshold you have to cross before the processor will drop down to base speed. I can confirm this 100% with the research I've done.
Well, now that we have found real TDP value of this CPU, your nonsense about comparing it with retail counterparts eventually found its ground. As I already posted 2667 v2 requires 134.6W to maintain 3.5GHz frequency on 8 cores under heavy AVX load, and your CPU will behave exactly similar - it will drop the multiplier to 35 consuming 135W of power (PL2, 'real' TDP, or whatever you want to call it). No benefit over 2667v2 and costs more. Yet unknown how it will behave on other vendor motherboard with no 'unlimited' tau
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
What does this even mean? Once again you are failing to take into account the fact that the 2673 v2 has an all core turbo of 3.6GHz (as does the 130w 2667 v2 and the fire alarm 150 watt 2687w v2). And the Power limit 2 on the 2673 v2 is rated at 137.5W with an unlimited time component. Meaning that it will not throttle back to "base clock" under any work load.
Holy cow! Are you telling us that your 2673v2 costs twice as much as 2667v2 with ONLY a ~0MHz benefit and 7.5W higher consumption (137.5w against 130W)? That's the most ridiculously priced CPU I've ever seen! :D
 

Storm-Chaser

Twin Turbo
Apr 16, 2020
151
25
28
Upstate NY
LOL. PL2=forever is a feature of motherboard vendors that doesn't grant Intel recommendations, not a particular CPU feature. And if your motherboard set tau=unlimited value automatically or allows you to set it manually doesn't mean this CPU will behave similarly in other motherboards such as Supermicro or Dell.
Wow, still pushing this flawed logic I see. if you want to keep going with this erroneous logic you will continue to dig your own grave, everyone is going to see it and it's pretty obvious its not motherboard related, revealing a total lack of research on your part. Because if I remember correctly you said on multiple occasions this OEM CPU was limited by processor TDP. Now that I have revealed that as being completely untrue, perhaps you should do a little more research before you post again. PL1 and PL2 are features of the CPU itself and have nothing to do with the motherboard or motherboard vendors, they are completely related to the the architecture of the turbo system created by Intel and by intel alone.

Still think PL1 and PL2 are motherboard vendor creations? See where I highlighted the part that says CPUID properties? Yeah, it's the subsect in AIDA64 that yields ID data related directly to the CPU itself.
1630765650486.png
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
PL1 and PL2 are features of the CPU itself and have nothing to do with the motherboard or motherboard vendors, they are completely related to the the architecture of the turbo system created by Intel and by intel alone.
Wrong again. It's motherboard feature. Intel tau for PL2 = 7.81ms
1630766975482.png
 

Whaaat

Active Member
Jan 31, 2020
301
157
43
Wow, still pushing this flawed logic I see. if you want to keep going with this erroneous logic you will continue to dig your own grave, everyone is going to see it and it's pretty obvious its not motherboard related, revealing a total lack of research on your part. Because if I remember correctly you said on multiple occasions this OEM CPU was limited by processor TDP. Now that I have revealed that as being completely untrue, perhaps you should do a little more research before you post again. PL1 and PL2 are features of the CPU itself and have nothing to do with the motherboard or motherboard vendors, they are completely related to the the architecture of the turbo system created by Intel and by intel alone.
Now that we see you are wrong again, stop writing this horsepoo and show us how much your 2673v2 consumes in Watts under Linx using HWinfo similar to this screen and we will decide if it worth the double price of 2667v2.