The IOQD BIOS feature controls the use of the processor bus’ command queue.
Normally, there are only two options available. Depending on the motherboard chipset, the options could be (1 and 4), (1 and 8) or (1 and 12).
The first queue depth option is always 1, which prevents the processor bus pipeline from queuing any outstanding commands. If selected, each command will only be issued after the processor has finished with the previous one.
Therefore, every command will incur the maximum amount of latency. This varies from 4 clock cycles for a 4-stage pipeline to 12 clock cycles for pipelines with 12 stages.
In most cases, it is highly recommended that you enable command queuing by selecting the option of 4 / 8 / 12 or in some cases, Enabled.
This allows the processor bus pipeline to mask its latency by queuing outstanding commands. You can expect a significant boost in performance with this feature enabled.
Interestingly, this IOQD feature can also be used as an aid in overclocking the processor. Although the queuing of commands brings with it a big boost in performance, it may also make the processor unstable at overclocked speeds. To overclock beyond what’s normally possible, you can try disabling command queuing.
But please note that the performance deficit associated with deeper pipelines (8 or 12 stages) may not be worth the increase in processor overclockability. This is because the deep processor bus pipelines have very long latencies.
If they are not masked by command queuing, the processor may be stalled so badly that you may end up with poorer performance even if you are able to further overclock the processor. So, it is recommended that you enable command queuing for deep pipelines, even if it means reduced overclockability.
IOQD : The Details
For greater performance at high clock speeds, motherboard chipsets now feature a pipelined processor bus. The multiple stages in this pipeline can also be used to queue up multiple commands to the processor.
This command queuing greatly improves performance because it effectively masks the latency of the processor bus. In optimal situations, the amount of latency between each succeeding command can be reduced to only a single clock cycle!
The IOQD BIOS feature controls the use of the processor bus’ command queue. Normally, there are only two options available.
Depending on the motherboard chipset, the options could be (1 and 4), (1 and 8) or (1 and 12). This is because the IOQD BIOS feature does not actually allow you to select the number of commands that can be queued.
It merely allows you to disable or enable the command queuing capability of the processor bus pipeline. This is because the number of commands that can be queued depends entirely on the number of stages in the pipeline.
As such, you can expect to see IOQD to be associated with options like Enabled and Disabled in some motherboards.
The first queue depth option is always 1, which prevents the processor bus pipeline from queuing any outstanding commands.
If selected, each command will only be issued after the processor has finished with the previous one. Therefore, every command will incur the maximum amount of latency. This varies from 4 clock cycles for a 4-stage pipeline to 12 clock cycles for pipelines with 12 stages.
As you can see, this reduces performance as the processor has to wait for each command to filter down the pipeline. The severity of the effect depends greatly on the depth of the pipeline. The deeper the pipeline, the greater the effect.
If the second queue depth option is 4, this means that the processor bus pipeline has 4 stages in it. Selecting this option allows the queuing of up to 4 commands in the pipeline. Each command can then be processed successively with a latency of only 1 clock cycle.
If the second queue depth option is 8, this means that the processor bus pipeline has 8 stages in it. Selecting this option allows the queuing of up to 8 commands in the pipeline. Each command can then be processed successively with a latency of only 1 clock cycle.
If the second queue depth option is 12, this means that the processor bus pipeline has 12 stages in it. Selecting this option allows the queuing of up to 12 commands in the pipeline. Each command can then be processed successively with a latency of only 1 clock cycle.
Please note that the latency of only 1 clock cycle is only possible if the pipeline is completely filled up. If the pipeline is only partially filled up, then the latency affecting one or more of the commands will be more than 1 clock cycle. Still, the average latency for each command will be much lower than it would be with command queuing disabled.
In most cases, it is highly recommended that you enable command queuing by selecting the option of 4 / 8 / 12 or in some cases, Enabled. This allows the processor bus pipeline to mask its latency by queuing outstanding commands. You can expect a significant boost in performance with this feature enabled.
Interestingly, this IOQD feature can also be used as an aid in overclocking the processor. Although the queuing of commands brings with it a big boost in performance, it may also make the processor unstable at overclocked speeds.
To overclock beyond what’s normally possible, you can try disabling command queuing. This may reduce performance but it will make the processor more stable and may allow it to be further overclocked.
But please note that the performance deficit associated with deeper pipelines (8 or 12 stages) may not be worth the increase in processor overclockability. This is because the deep processor bus pipelines have very long latencies.
If they are not masked by command queuing, the processor may be stalled so badly that you may end up with poorer performance even if you are able to further overclock the processor. So, it is recommended that you enable command queuing for deep pipelines, even if it means reduced overclockability.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
AMD just introduced nineteen EPYC 7002 series processors, with 8 to 64 cores and up to 256 MB of L3 cache!
To make it easier for you to differentiate them, we compared their key specifications and prices.
AMD EPYC 7002 Series
Based on the AMD Zen 2 microarchitecture, the AMD EPYC 7002 series is the 2nd Gen EPYC family of server processors.
Built on the latest 7 nm process technology, AMD claims they will offer up to 90% better integer performance and up to 79% better floating-point performance, than the competing Intel Xeon Platinum 8280 processor.
On top of significantly better performance per socket, they also come with hardware memory encryption, and a dedicated security processor.
AMD EPYC 7002 Series Is Already Changing The Industry
AMD appeared to have shipped the EPYC 7002 series processors early to Google, where they were deployed in production servers for their internal datacenter infrastructure.
Google also plans to use the EPYC 7002 processors in new general-purpose machines that are part of the Google Cloud Compute Engine.
Twitter has also announced that they are already using the EPYC 7002 processors to reduce their datacenter TCO (total cost of ownership) by 25%.
For those who are looking to switch, HPE and Lenovo announced that they can provide datacenter solutions based on the EPYC 7002 processors.
[adrotate group=”1″]
AMD EPYC 7002 Specifications + Price Compared!
For your convenience, we compared the specifications and prices of the AMD EPYC 7002 models!
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
Leveraging their new Zen 2 microarchitecture and 7 nm process technology, AMD just introduced their 2nd Gen EPYC processors.
Designed to challenge Intel Xeon in the enterprise, cloud and HPC markets, the 2nd Gen EPYC processors promise to deliver “record-setting performance“, while reducing TCO (Total Cost of Ownership) by up to 50%.
Here is everything you need to know about the new 2nd Gen EPYC processors… summarised!
The Official 2nd Gen EPYC Product Presentation Summary
Let’s start with a quick 7.5 minute summary of the 2nd Gen EPYC product presentations by Dr. Lisa Su, Mark Papermaster and Forrest Norrod!
Now, let’s take a look at its key features and specifications!
AMD Infinity Architecture Explained
The AMD Infinity Architecture is a fancy name for their new modular chiplet-based design. It allows them to combine up to eight processor dies with a single I/O die on the same package, faster and at lower cost.
The processor dies are fabricated with the industry-leading 7 nm process technology for best performance at lowest power consumption, and thermal output.
The I/O die, on the other hand, can be fabricated on the much cheaper 14 nm process technology, with a much higher yield.
2nd Gen EPYC Is Built On 7nm
The 2nd Gen EPYC processor cores are fabricated on the 7nm process technology. This allows AMD to fit more transistors into a smaller space.
By doubling the transistor density, coupled with microarchitectural optimisations, the 2nd Gen EPYC delivers 4X the floating point performance of the 1st Gen EPYC processors.
The smaller process also increases energy efficiency, reducing both power consumption and heat output. According to AMD, 2nd Gen EPYC will use half the power consumption as the 1st Gen EPYC at the same performance level.
Industry-Leading Performance
AMD claims they will offer up to 90% better integer performance and up to 79% better floating-point performance, than the competing Intel Xeon Platinum 8280 processor.
On top of significantly better performance per socket, they also come with hardware memory encryption, and a dedicated security processor.
Baked-In Security On Multiple Levels
The 2nd Gen EPYC processors are built-in with multiple levels of security features, to harden it against cyberattacks.
They have a secure root of trust designed to validate the initial BIOS boot without corruption.
In virtualised environments, you can use it to cryptographically check that your entire software stack is booted without corruption.
They have memory encryption engines built into their memory channels to hardware-encrypt data in the memory, preventing cold boot attacks.
In the 2nd Gen EPYC, every virtual machine is now encrypted with one of up to 509 unique encryption keys known only to the processor.
This protects your data even if a malicious VM finds its way into your virtual machine memory, or if a compromised hypervisor gains access into a guest VM.
2nd Gen EPYC Is PCI Express Gen 4 Ready!
Like the 3rd Gen Ryzen processors, the 2nd Gen EPYC is PCI Express Gen 4 ready.
PCIe 4.0 doubles the bandwidth over PCIe 3.0, and every EPYC processor has 128 lanes to tie together HPC clusters, or connect to GPU accelerators and NVMe drives.
[adrotate group=”1″]
2nd Gen EPYC Model, Specifications + Price Summary
For your convenience, we summarised the specifications and prices of the 2nd Gen EPYC models!
64-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7742
64 / 128
2.25 GHz
3.4 GHz
256 MB
225 W
$6,950
EPYC 7702
64 / 128
2.0 GHz
3.35 GHz
256 MB
200 W
$6,450
EPYC 7702P
64 / 128
2.0 GHz
3.35 GHz
256 MB
200 W
$4,425
48-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7642
48 / 96
2.3 GHz
3.3 GHz
256 MB
225 W
$4,775
EPYC 7552
48 / 96
2.2 GHz
3.3 GHz
192 MB
200 W
$4,025
32-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7542
32 / 64
2.9 GHz
3.4 GHz
128 MB
225 W
$3,400
EPYC 7502
32 / 64
2.5 GHz
3.35 GHz
128 MB
180 W
$2,600
EPYC 7502P
32 / 64
2.5 GHz
3.35 GHz
128 MB
180 W
$2,300
EPYC 7452
32 / 64
2.35 GHz
3.35 GHz
128 MB
155 W
$2,205
24-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7402
24 / 48
2.8 GHz
3.35 GHz
128 MB
180 W
$1,783
EPYC 7402P
24 / 48
2.8 GHz
3.35 GHz
128 MB
180 W
$1,250
EPYC 7352
24 / 48
2.3 GHz
3.2 GHz
128 MB
155 W
$1,350
16-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7302
16 / 32
3.0 GHz
3.3 GHz
128 MB
155 W
$978
EPYC 7302P
16 / 32
3.0 GHz
3.3 GHz
128 MB
155 W
$825
EPYC 7282
16 / 32
2.8 GHz
3.2 GHz
64 MB
120 W
$650
12-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7272
12 / 24
2.9 GHz
3.2 GHz
64 MB
120 W
$625
12-Core Models
Cores /
Threads
Base Clock
Boost Clock
L3 Cache
TDP
1K Price
EPYC 7262
8 / 16
3.2 GHz
3.4 GHz
128 MB
155 W
$575
EPYC 7252
8 / 16
3.1 GHz
3.2 GHz
64 MB
120 W
$475
EPYC 7232P
8 / 16
3.1 GHz
3.2 GHz
32 MB
120 W
$450
2nd Gen EPYC Is Already Changing The Industry
AMD appeared to have shipped the 2nd Gen EPYC processors early to Google, where they were deployed in production servers for their internal datacenter infrastructure.
Google also plans to use the 2nd Gen EPYC processors in new general-purpose machines that are part of the Google Cloud Compute Engine.
Twitter has also announced that they are already using the 2nd Gen EPYC processors to reduce their datacenter TCO (total cost of ownership) by 25%.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
One of the biggest revelations at the AMD Computex 2018 press conference is how well along AMD is with their 7nm efforts. Everything appears to be chugging along as planned, and AMD shared not only details about their 7nm EPYC processor, they also showed off an engineering sample!
The 7nm EPYC Revealed!
At AMD Computex 2018, Dr. Lisa Su announced that AMD has decided to introduce 7nm Zen 2 processor, in the form of the 7nm EPYC processors!
Here are the key points in Dr. Lisa Su’s presentation :
AMD will go straight to 7nm Zen 2 cores for the next-generation EPYC processors.
7nm Zen 2 silicon are in AMD labs, and “it’s looking good!“
AMD will be sampling 7nm EPYC processors in the second half of 2018., with a launch in 2019.
7nm EPYC processors will continue to use Socket SP3, and will be “drop-in” replacements for the first-generation EPYC processors.
First Look At 7nm EPYC + 7nm Vega!
In this video, Dr. Lisa Su shows off engineering samples of the 7nm EPYC processor (on the left), and the 7nm Vega GPU (on the right).
First Examples Of 7nm EPYC Servers
AMD also showed examples of a single CPU server, and a dual CPU server, with 7nm EPYC processors.
As AMD hasn’t actually produced any 7nm EPYC thus far (sampling only starts in the second half of 2018), these are just dummy CPU mockups. The real purpose is to demonstrate that they will be pin-compatible drop-in replacements for existing EPYC servers… with a BIOS update, of course!
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
The speculative execution CPU bug that literally kneecapped Intel, also affects many AMD and ARM processors. This means BILLIONS of CPUs around the world, including those powering smartphones, are affected by Meltdown and/or Spectre.
Our article Everything On The Meltdown + Spectre CPU Flaws! summarises the key details of the speculative execution bug, and what we can do about it. This guide is to help those who want a full list of affected CPUs. Because we intend this to be an exhaustive list, we split it into multiple sections.
Article Update History
Click here for the Article Update History
Updated @ 2018-03-07 : Added a new list of 5 IBM z/Architecture CPUs. Added a new list of 22 VIA desktop and mobile CPUs. Added 1 ARM mobile CPU, 1 Intel server CPU, and 1 Intel mobile CPU. Also added 20 mobile SoCs, 9 digital TV or media player SoCs, and 43 industrial SoCs.
Updated @ 2018-02-15 : Added 96 Intel server CPUs, 91 Intel desktop CPUs, and 127 Intel mobile CPUs.
Updated @ 2018-02-07 : Added 128 AMD server CPUs, 11 AMD workstation CPUs, 128 AMD desktop CPUs, and 59 AMD mobile CPUs.
Updated @ 2018-02-02 : Added 11 Intel server CPUs, 96 AMD server CPUs, 168 AMD desktop CPUs, 77 AMD mobile CPUs, 10 IBM POWER CPUs, 9 HiSilicon Kirin mobile SoCs, 10 MediaTek mobile SOCs, 4 MediaTek digital TV SoCs, and 6 NVIDIA devices to the lists of vulnerable CPUs.
Updated @ 2018-01-14 : Added 416 Intel server CPUs, 8 Intel desktop CPUs, and 29 Intel mobile CPUs to the lists of vulnerable CPUs. Added a new list of 51 Intel mobile SoCs.
Updated @ 2018-01-12 : Added 71 AMD server CPUs, 71 AMD desktop CPUs, 29 AMD mobile CPUs and 3 AMD server SoCs based on a vulnerable ARM CPU. Also added a table summarising the number of vulnerable processors.
Updated @ 2018-01-11 : Added 18 Intel desktop CPUs and 165 Intel server / workstation CPUs. Also added a list of vulnerable Apple iOS devices, and expanded the list of vulnerable mobile SoCs used by smartphones.
Originally posted @ 2018-01-08
What Are Meltdown And Spectre?
Meltdown and Spectre are two exploits that take advantage of three variants of the speculative execution bug that affects billions of CPUs around the world.
The Spectre exploit targeted Variants 1 and 2, while the Meltdown exploit targets Variant 3, of the CPU bug.
The CPUs Vulnerable To Meltdown / Spectre Updated!
For easy reference, we divided the affected CPUs by Company (arranged ALPHABETICALLY – no conspiracy, we promise), and subsequently by Segment (Workstation / Desktop / Mobile), or affected variants.
As of Revision 8.0, we believe we have covered all of the affected AMD, Apple, ARM, IBM, Intel and VIA CPUs. But we will add more CPUs (and devices) as and when they’re noted to be vulnerable to the Meltdown and Spectre exploits.
Note : It’s arguable that all CPUs that uses speculative execution to any degree are potentially vulnerable to Meltdown or Spectre or a future exploit. We will only focus on CPUs that are confirmed to be vulnerable to Meltdown or Spectre.
Vulnerable CPUs By The Numbers Updated!
Here is a quick summary of the number of CPUs vulnerable to Meltdown or Spectre, according to the company, and the type of processor.
Company
Spectre 1
Spectre 2
Meltdown
AMD
295 Server CPUs
42 Workstation CPUs
396 Desktop CPUs
208 Mobile CPUs
295 Server CPUs
42 Workstation CPUs
396 Desktop CPUs
208 Mobile CPUs
None
Apple
13 Mobile SoCs
13 Mobile SoCs
13 Mobile SoCs
ARM
10 Mobile CPUs
3 Server SoCs
10 Mobile CPUs
3 Server SoCs
4 Mobile CPUs
3 Server SoCs
IBM
5 z/Architecture CPUs
10 POWER CPUs
5 z/Architecture CPUs
10 POWER CPUs
5 z/Architecture CPUs
10 POWER CPUs
Intel
733 Server / Workstation CPUs
443 Desktop CPUs
584 Mobile CPUs
51 Mobile SoCs
733 Server / Workstation CPUs
443 Desktop CPUs
584 Mobile CPUs
51 Mobile SoCs
733 Server / Workstation CPUs
443 Desktop CPUs
584 Mobile CPUs
51 Mobile SoCs
Affected Variants :AMD CPUs are affected by both Variants 1 and 2 of the speculative execution CPU bug. Colloquially, many people refer to them as Spectre 1 and Spectre 2.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
The AMD Workstation CPUs Vulnerable To Spectre
Affected Variants :AMD CPUs are affected by both Variants 1 and 2 of the speculative execution CPU bug. Colloquially, many people refer to them as Spectre 1 and Spectre 2. They are not vulnerable to Meltdown.
AMD Summit Ridge (2017)
AMD Ryzen Threadripper 1950X
AMD Ryzen Threadripper 1920X
AMD Ryzen Threadripper 1900X
AMD Vishera (2012)
AMD FX-9590
AMD FX-9370
AMD FX-8370E
AMD FX-8370
AMD FX-8350
AMD FX-8320E
AMD FX-8320
AMD FX-8310
AMD FX-8300
AMD FX-6350
AMD FX-6300
AMD FX-6200
AMD FX-4350
AMD FX-4320
AMD FX-4300
AMD Zambezi (2011)
AMD FX-8170
AMD FX-8150
AMD FX-8140
AMD FX-8120
AMD FX-8100
AMD FX-6130
AMD FX-6120
AMD FX-6100
AMD FX-4170
AMD FX-4150
AMD FX-4130
AMD FX-4120
AMD FX-4100
AMD Windsor (2006)
AMD Athlon 64 FX-74
AMD Athlon 64 FX-72
AMD Athlon 64 FX-70
AMD Athlon 64 FX-62
AMD Toledo (2005)
AMD Athlon 64 FX-60
AMD San Diego (2005)
AMD Athlon 64 FX-57
AMD Athlon 64 FX-55
AMD Clawhammer (2004)
AMD Athlon 64 FX-55
AMD Athlon 64 FX-53
AMD Sledgehammer (2003)
AMD Athlon 64 FX-53
AMD Athlon 64 FX-51
[adrotate group=”1″]
AMD Desktop CPUs Vulnerable To Spectre
Affected Variants :AMD CPUs are affected by both Variants 1 and 2 of the speculative execution CPU bug. Colloquially, many people refer to them as Spectre 1 and Spectre 2. They are not vulnerable to Meltdown.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
AMD Mobile CPUs Vulnerable To Spectre
Affected Variants :AMD CPUs are affected by both Variants 1 and 2 of the speculative execution CPU bug. Colloquially, many people refer to them as Spectre 1 and Spectre 2. They are not vulnerable to Meltdown.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
The Apple CPUs Vulnerable To Meltdown / Spectre
Apple makes custom processors based on the ARM microarchitecture. They have not released specific information on which of their processors are affected by which exploit, but this is what we know so far.
Affected Variants : Apple only issued a general notice that their processors are affected by both Meltdown and Spectre, not the specific variants.
Apple A4
Apple A5
Apple A5X
Apple A6
Apple A6X
Apple A7
Apple A8
Apple A8X
Apple A9
Apple A9X
Apple A10 Fusion
Apple A10X Fusion
Apple A11 Bionic
Vulnerable iOS or tvOS Devices : Apple was vague about the iOS devices that were affected, but based on the affected CPU cores, here are the iOS devices that are vulnerable to Meltdown and Spectre :
Apple TV 2nd Generation, 3rd Generation, 4th Generation and 5th Generation
The ARM CPUs Vulnerable To Meltdown / Spectre
ARM CPUs Vulnerable To All Three Variants
Affected Variants :Variants 1 and 2, and either Variant 3 or Variant 3a, of the speculative execution CPU bug. They are vulnerable to Meltdown and both variants of Spectre.
ARM Cortex-A75
ARM Cortex-A72
ARM Cortex-A57
ARM Cortex-A15
Mobile SoCs Using These ARM CPUs (Not Exhaustive)
HiSilicon Kirin 955
HiSilicon Kirin 950
HiSilicon Kirin 928
HiSilicon Kirin 925
HiSilicon Kirin 920
MediaTek Helio X27 (MT6797X)
MediaTek Helio X25 (MT6797T)
MediaTek Helio X23 (MT6707D)
MediaTek Helio X20 (MT6797)
MediaTek MT8173
MediaTek MT8135 / MT8135V
MediaTek MT6795
NVIDIA Tegra X2
NVIDIA Tegra X1
NVIDIA Tegra K1
NVIDIA Tegra 4
Qualcomm Snapdragon 845
Qualcomm Snapdragon 810 / 808
Qualcomm Snapdragon 670
Qualcomm Snapdragon 653 / 652 / 650
Qualcomm Snapdragon 640
Samsung Exynos 7420
Samsung Exynos 5800
Samsung Exynos 5433
Samsung Exynos 5422 / 5420
Samsung Exynos 5410
Samsung Exynos 5260
Samsung Exynos 5250
Samsung Exynos 5 Dual (Exynos 5250)
AMD Server SoCs Using These ARM CPUs
AMD Opteron A1170
AMD Opteron A1150
AMD Opteron A1120
NVIDIA Devices Using These ARM CPUs (Not Exhaustive)
NVIDIA SHIELD TV (ARM Cortex-A57)
NVIDIA SHIELD Tablet (ARM Cortex-A15)
NVIDIA Jetson TX2 (ARM Cortex-A57)
NVIDIA Jetson TX1 (ARM Cortex-A57)
NVIDIA Jetson TK1 (ARM Cortex-A15)
NVIDIA Jetson Tegra K1 (ARM Cortex-A15)
Digital TV / Media Player SoCs Using These ARM CPUs (Not Exhaustive)
Rockchip RK3399
Industrial SoCs Using These ARM CPUs (Not Exhaustive)
Embedded Computers Using These ARM CPUs (Not Exhaustive)
VIA VAB-1000
VIA VAB-820 / VAB-800
VIA VAB-630 / VAB-600
VIA ALTA DS
VIA QSM-8Q60
VIA SOM-6X50
VIA VTS-8589
IBM POWER CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These IBM POWER CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
IBM POWER4
IBM POWER4+
IBM POWER5
IBM POWER5+
IBM POWER6
IBM POWER6+
IBM POWER7
IBM POWER7+
IBM POWER8
– including IBM Murano, IBM Turismo, PowerCore CP1
IBM POWER8 with NVLink / POWER8+
IBM POWER9
– IBM Nimbus, IBM Cumulus
IBM z/Architecture CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These IBM z/Architecture CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
IBM z14
IBM z13
IBM zEC12
IBM z196
IBM z10
[adrotate group=”1″]
Intel UMPC / Smartphone SoCs Vulnerable To Meltdown + Spectre
Affected Variants : These Intel SoCs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
Intel Server / Workstation CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These Intel CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
Intel Desktop CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These Intel CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
Intel Mobile CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These Intel CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
VIA Desktop CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These VIA CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
VIA Nano QuadCore (2011)
VIA Nano QuadCore L4800E
VIA Nano QuadCore L4700E
VIA Nano QuadCore L4650E
VIA Nano Dual Core 2011)
VIA Nano X2 E L4350E
VIA Nano X2 E L4350E
VIA Nano 3000 Series (2009)
VIA Nano L3600
VIA Nano L3050
VIA Nano L3025
VIA Nano 2000 Series (2008)
VIA Nano L2200
VIA Nano L2100
VIA Mobile CPUs Vulnerable To Meltdown + Spectre
Affected Variants : These VIA CPUs are affected by all three variants of the speculative execution CPU bug. They are vulnerable to the Meltdown and both Spectre exploits.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
It is common knowledge that the latest AMD Ryzen 7, Ryzen 5 and Ryzen 3 are vulnerable to the Spectre exploits, and so are the older AMD processors based on their Bulldozer microarchitecture. But what about the AMD K10 and K8 processors? Today, we can confirm that AMD K10 and K8 processors from as far back as 2003 are vulnerable to Spectre.
AMD K10 And K8 Processors Are Also Vulnerable To Spectre
While it was common knowledge that the latest AMD Ryzen 7, Ryzen 5 and Ryzen 3, as well as the Bulldozer-based APUs are vulnerable to Spectre, no one knew for sure whether the older AMD K10 and K8 processors were also affected. After all, they were based on a different microarchitecture.
That was until they revealed their Spectre mitigation strategies. Hidden in their whitepaper were references to AMD Family 10h (also known as AMD K10), as well as AMD Family 0Fh (also known as AMD K8).
So we now know, for sure, that the AMD K10 and AMD K8 processors are also vulnerable to Spectre.
What Are The Affected AMD K10 & K8 Processors?
The AMD K10 family alone adds 96 AMD server CPUs, 168 AMD desktop CPUs, and 77 AMD mobile CPUs to the list of vulnerable processors.
The AMD K8 family adds a further 128 AMD server CPUs, 137 AMD desktop CPUs, and 57 AMD mobile CPUs to the list.
For your convenience, we included the entire list of vulnerable AMD K10 and AMD K8 processors, split into three lists :
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
AMD just launched the AMD EPYC 7000 series processors. Formerly known as AMD Naples, it boasts up to 32 AMD Zen processor cores, and challenges the Intel Xeon’s dominance in the datacenter. After a sneak peek at their Computex 2017 press conference, AMD finally revealed the EPYC 7000 series processors. Let’s take a look!
The AMD EPYC 7000 Series Processor
Like Ryzen, the EPYC processor is based on their new Zen microarchitecture. Designed for servers, the EPYC processor offers up to 32 processor cores, an integrated high-speed DDR4 memory controller and a new high-speed coherent interconnect. AMD now reveals that it also comes with an embedded security subsystem.
Here is a summary of the EPYC processor’s key features :
Supports up to 32 AMD Zen cores., each capable of handling 2 simultaneous threads.
[adrotate group=”2″]Integrated DDR4 memory controller with 8 memory channels, each supporting two DIMMs. That’s a total of 16 DIMMs per processor socket, allowing for up to 2 TB of memory. A 2-socket server will support up to 32 DIMMs, with a total memory capacity of 4 TB.
The EPYC processor is fully integrated with high-speed I/O including 128 PCIe 3.0 lanes.
Because the EPYC is a complete SoC, there is no need for a separate chipset, reducing cost and power consumption.
In a 2-socket system, two EPYC processors communicate with each other via the AMD Infinity Fabric coherent interconnect. Even though each EPYC processor has 128 PCIe 3.0 lanes, it is limited to 64 PCIe 3.0 lanes in a 2-socket design. The 2P EPYC system will have a total of 128 PCI 3.0 lanes, not 256.
In addition, AMD revealed the first EPYC family of processors – the EPYC 7000 Series, with nine processors divided into four segments. They also introduced three EPYC processor variants designed for single socket servers.
EPYC 7601, 7551, 7551P and 7501 processors have 32 cores and a peak TDP of 170-180 W.
EPYC 7451, 7401 and 7401P processors have 24 cores and a peak TDP of 170-180 W.
EPYC 7351, 7351P, 7301 and 7281 processors have 16 cores and a peak TDP of 170 W.
EPYC 7251 processor has 8 cores and a peak TDP of 120 W.
The AMD EPYC 7000 Series Presentation
If you have the time, check out this 103-minute presentation on the EPYC 7000 Series processors by Dr. Lisa Su and Forrest Norrod from AMD, as well as their partners. Dr. Lisa also comes out at the end to reveal the AMD Radeon Instinct MI25 accelerator, that we revealed several months earlier.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
EPYC 7000 Series Processor Performance
AMD also revealed their internal benchmarks for the EPYC 7000 series processors :[adrotate group=”2″]
Single Socket AMD EPYC 7601 Server
SPECiut_rate2006 : 1200
SPECfp_rate2006 : 943
Two Socket AMD EPYC 7601 Server
SPECiut_rate2006 : 1390
SPECfp_rate2006 : 1330
AMD compared the two processor performance of the EPYC 7601, showing that it beat the Intel Xeon E5-2699A v4 in integer compute by 47%, and floating point compute by 75%.
Price Performance Advantage
AMD also took pains to point out that the price performance advantage the EPYC 7000 Series processors have over their Intel Xeon rivals, in both the 2-socket and single-socket segments.
Data Security Advantage
Other than raw computing performance and support for a truckload of memory, EPYC processors also offer an integrated hardware security subsystem. This allows the EPYC processors to deliver full memory encryption and secure multi-tenancy for data domain with no application impact.
Compatibility & Support
Finally, AMD wants everyone to know that the EPYC is an x86 processor and readily supports all x86 applications in the market. They have also performed extensive tests with industry partners like Microsoft, VMware and Red Hat.
They have also lined up considerable support from their server partners to deliver AMD EPYC solutions to customers worldwide.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
AMD EPYC 7000 Series Processor Press Release
AUSTIN, Texas — June 20, 2017 — AMD (NASDAQ: AMD), and a global ecosystem of server partners, today marked a new era in the datacenter with the launch of AMD EPYC 7000 series high-performance datacenter processors. AMD was joined by multiple customers and partners at the global launch event in presenting a wide array of systems, performance demonstrations, and customer testimonials. The innovative, record-setting EPYC design, with up to 32 high-performance “Zen” cores and an unparalleled feature set, delivers greater performance than the competition across a full range of integer, floating point, memory bandwidth, and I/O benchmarks and workloads.
“With our EPYC family of processors, AMD is delivering industry-leading performance on critical enterprise, cloud, and machine intelligence workloads,” said Lisa Su, president and CEO, AMD. “EPYC processors offer uncompromising performance for single-socket systems while scaling dual-socket server performance to new heights, outperforming the competition at every price point. We are proud to bring choice and innovation back to the datacenter with the strong support of our global ecosystem partners.”
The world’s largest server manufacturers introduced products based on EPYC 7000-series processors at today’s launch, including HPE, Dell, ASUS, Gigabyte, Inventec, Lenovo, Sugon, Supermicro, Tyan, and Wistron. Primary hypervisor and server operating system providers Microsoft, Red Hat, and VMware showcased optimized support for EPYC, while key server hardware ecosystem partners Mellanox, Samsung Electronics, and Xilinx were also featured in EPYC-optimized platforms.
Cloud Datacenter and Enterprise Customers
Datacenter and cloud service providers also welcomed EPYC to the market today. Members of the “Super 7” datacenter services providers, including Baidu and Microsoft Azure, as well as 1&1, Bloomberg, Dropbox and LexisNexis, all voiced their support at launch.
Record-Setting EPYC Performance
The excitement around EPYC is driven by multiple record-setting server benchmarks achieved by EPYC-powered one-socket and two-socket systems.
AMD EPYC processors set several performance records, including:
Two-Socket Server
EPYC 7601-based system scored 2360 on SPECint_rate2006, higher than any other two-socket system score
One-Socket Server
EPYC 7601-based system scored 1200 on SPECint_rate2006, higher than any other mainstream one-socket x86-based system score
EPYC 7601-based system scored 943 on SPECfp_rate2006, higher than any other one-socket system score
All EPYC processors combine innovative security features, enterprise class reliability, and support a full feature-set. An EPYC 7601 CPU-based one-socket system shifts expectations for single socket server performance, helping lower total-cost-of-ownership (TCO), providing up to 20% CapEx savings compared to the Intel Xeon E5-2660 v4-based two-socket system. At every targeted price point for two-socket processors, EPYC outperforms the competition, with up to 70% more performance in the eight hundred dollar price band and up to 47% more performance at the high-end of the market of four thousand dollars or more.
EPYC Product Overview
A highly scalable System on Chip (SoC) design ranging from 8-core to 32-core, supporting two high-performance threads per core.
Industry-leading memory bandwidth across the line-up, with 8 channels of memory on every EPYC device. In a two-socket server, support for up to 32 DIMMS of DDR4 on 16 memory channels, delivering up to 4 terabytes of total memory capacity.
Unprecedented support for integrated, high-speed I/O with 128 lanes of PCIe 3 on every product
A highly-optimized cache structure for high-performance, energy efficient compute
AMD Infinity Fabric coherent interconnect linking EPYC CPUs in a two-socket system
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
Like Ryzen, the AMD Naples CPU is based on their brand-new Zen microarchitecture. Designed for servers, the Naples CPU offers up to 32 processor cores, an integrated high-speed DDR4 memory controller and a new high-speed coherent interconnect.
These four slides summarises the Naples CPU’s key features :
32 AMD Zen cores., each capable of handling 2 simultaneous threads.
Integrated DDR4 memory controller with 8 memory channels, each supporting two DIMMs. That’s a total of 16 DIMMs per processor socket, allowing for up to 2 TB of memory. A 2-socket server will support up to 32 DIMMs, with a total memory capacity of 4 TB.
The Naples CPU is fully integrated with high-speed I/O including 128 PCIe 3.0 lanes.
Because the Naples is a complete SoC, there is no need for a separate chipset, reducing cost and power consumption.
In a 2-socket system, two Naples CPUs will communicate with each other via the AMD Infinity Fabric coherent interconnect. Even though each Naples CPU technically supports 128 PCIe 3.0 lanes, it is limited to 64 PCIe 3.0 lanes in a 2-socket design.
A 2-socket Naples server would offer 64 processor cores that can handle 128 simultaneous cores, and up to 4 TB of DDR4 memory (32 x 128 GB DIMMs). Just how fast is such a server? Let’s find out…
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
The AMD Naples CPU Performance
AMD benchmarked the performance of a 2-socket Naples server against a 2-socket Intel Xeon E5-2699A v4 server, with the following specifications. The workload was Seismic Analysis, involving iterations of 3D wave equations.
When they were compared with similar number of cores (44), the Naples CPU was almost twice as fast as the Intel Xeon E5-2699A v4 CPU.
But when they unleashed all 64-cores in that 2-socket Naples server, it is now 2.5X faster than the Intel Xeon E5-2699A v4.
In this final test, they tested both servers with a 4x larger dataset – 4 billion samples, instead of just 1 billion samples. Due to its limited memory channels, the Intel Xeon E5-2699A v4 is unable to support enough memory to load the dataset.
[adrotate banner=”5″]
AMD Naples + Radeon Instinct
Naples is more than just a server CPU to AMD. Its integrated support for 128 PCIe 3.0 lanes allows a single Naples CPU to support up to 4 AMD Radeon Instinct accelerators.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
The AMD Naples Pricing & Availability
You may wonder why AMD compared the Naples CPU against the 22-core Intel Xeon E5-2699A v4, instead of a more powerful processor like the Intel Xeon E7-8893 v4. We suspect it is because they are going to price the Naples CPU against the Xeon E5-2699A v4, which costs US$ 4,938.
Imagine how well it will sell if it’s priced at the same or slightly lower price point than the Xeon E5-2699A v4, while offering 45% more cores, 60% more I/O capacity and 122% more memory bandwidth. If we have to guess, we would think that AMD will finally price the Naples CPU at around $4,000 – a 20% discount on the Xeon E5-2699A v4.
More importantly, AMD appears to be on track for a Q2 2017 launch for the Naples CPU. We will keep you updated!
[adrotate group=”1″]
The AMD Naples Presentation Slides
Here is the complete set of slides used by Forrest Norrod, SVP and GM of the AMD Enterprise, Embedded and Semi-Custom Business Group.
If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!
