{"id":92,"date":"2023-12-17T15:14:45","date_gmt":"2023-12-17T15:14:45","guid":{"rendered":"https:\/\/mvermeulen.org\/perf\/?p=92"},"modified":"2023-12-17T19:31:27","modified_gmt":"2023-12-17T19:31:27","slug":"new-ryzen-7840-machine","status":"publish","type":"post","link":"https:\/\/mvermeulen.org\/perf\/2023\/12\/17\/new-ryzen-7840-machine\/","title":{"rendered":"New Ryzen 7840 machine"},"content":{"rendered":"\n

I have set up a new AMD performance machine for experiments. The processors is a Ryzen 7840<\/a> (Phoenix) in a Beelink SER7<\/a> mini-PC.<\/p>\n\n\n\n

Following are some of the major parameters. This comparison is with Intel i5-13500H which will be my Intel comparison microprocessor.<\/p>\n\n\n\n\n\n\n\n\t\n\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t\n\t
Item<\/th>Ryzen 7840HS<\/th>i5-13500H<\/th>Notes<\/th>\n<\/tr>\n<\/thead>\n
Architecture<\/td>Zen4<\/td>Raptor Lake<\/td><\/td>\n<\/tr>\n
Cores<\/td>8<\/td>12
\n4 performance (raptor cove)
\n8 efficiency (gracemont)<\/td>		<\/td>\n<\/tr>\n
Threads<\/td>16<\/td>16<\/td><\/td>\n<\/tr>\n
Base Clock<\/td>3.8 GHz<\/td>2.6 GHz, 1.9 GHz<\/td><\/td>\n<\/tr>\n
Boost Clock<\/td>5.1 GHz<\/td>4.7 GHz, 3.5 GHz<\/td><\/td>\n<\/tr>\n
TDP<\/td>35-45W<\/td>45W-95W<\/td>Set by vendor<\/td>\n<\/tr>\n
Memory<\/td>32 GB (2 x 16 GiB)
\n
\nDDR5 - 5600
\n
\n2 Memory Channels<\/td>		16 GB
\n
\nDDR4 - 3200
\n
\n2 Memory Channels<\/td>		Check BIOS for actual speed<\/td>\n<\/tr>\n
Stream<\/td>Copy: 71400 MB\/s
\nScale: 70300 MB\/s
\nAdd: 73600 MB\/s
\nTriad: 73000 MB\/s<\/td>		Copy: 39200 MB\/s
\nScale: 39100 MB\/s
\nAdd: 40100 MB\/s
\nTriad: 40000 MB\/s<\/td>		Measured<\/td>\n<\/tr>\n
Cache<\/td>L1 - 32kB, 8 way, 4 clocks
\n
\nL2 - 1 MB, 8-way, 14 clocks
\n
\nL3 - 16MB, 24 way, 47 clocks<\/td>		L1 - 48 kB, 12-way\/8-way, 3\/5 clocks
\n
\nL2 - 1 MB, 10-way\/16-way, 15-20 clocks
\n
\nL3 - 18 MB, 10-way, 65-20 clocks<\/td>		Agner Fog architecture document and likwid-topology<\/td>\n<\/tr>\n
lmbench<\/td>L1 - 0.8 ns
\nL2 - 3 ns
\nL3 - 8 ns<\/td>		L1 - 1.3 ns, 1.0 ns
\nL2 - 4.4 ns, 8ns
\nL3 - 12 ns, 19ns<\/td>		Measured in Nanoseconds<\/td>\n<\/tr>\n
Graphics<\/td>Radeon 780M
\n
\n12 cores
\n
\n2700 MHz<\/td>		Intel Iris Xe<\/td><\/td>\n<\/tr>\n
Phoronix stream<\/td>Average: 40604 MB\/s<\/td>Average: 35422 MB\/s<\/td>1.15x ratio smaller than optimized compiler results above<\/td>\n<\/tr>\n
Phoronix coremark<\/td>Average 464076 Iterations\/second<\/td>Average 388569 Iterations\/second<\/td>1.19 ratio<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n\n

Following is the topology shown by likwid-topology<\/p>\n\n\n\n

--------------------------------------------------------------------------------\nCPU name:\tAMD Ryzen 7 7840HS w\/ Radeon 780M Graphics     \nCPU type:\tnil\nCPU stepping:\t1\n********************************************************************************\nHardware Thread Topology\n********************************************************************************\nSockets:\t\t1\nCores per socket:\t8\nThreads per core:\t2\n--------------------------------------------------------------------------------\nHWThread\tThread\t\tCore\t\tSocket\t\tAvailable\n0\t\t0\t\t0\t\t0\t\t*\n1\t\t0\t\t1\t\t0\t\t*\n2\t\t0\t\t2\t\t0\t\t*\n3\t\t0\t\t3\t\t0\t\t*\n4\t\t0\t\t4\t\t0\t\t*\n5\t\t0\t\t5\t\t0\t\t*\n6\t\t0\t\t6\t\t0\t\t*\n7\t\t0\t\t7\t\t0\t\t*\n8\t\t1\t\t0\t\t0\t\t*\n9\t\t1\t\t1\t\t0\t\t*\n10\t\t1\t\t2\t\t0\t\t*\n11\t\t1\t\t3\t\t0\t\t*\n12\t\t1\t\t4\t\t0\t\t*\n13\t\t1\t\t5\t\t0\t\t*\n14\t\t1\t\t6\t\t0\t\t*\n15\t\t1\t\t7\t\t0\t\t*\n--------------------------------------------------------------------------------\nSocket 0:\t\t( 0 8 1 9 2 10 3 11 4 12 5 13 6 14 7 15 )\n--------------------------------------------------------------------------------\n********************************************************************************\nCache Topology\n********************************************************************************\nLevel:\t\t\t1\nSize:\t\t\t32 kB\nCache groups:\t\t( 0 8 ) ( 1 9 ) ( 2 10 ) ( 3 11 ) ( 4 12 ) ( 5 13 ) ( 6 14 ) ( 7 15 )\n--------------------------------------------------------------------------------\nLevel:\t\t\t2\nSize:\t\t\t1 MB\nCache groups:\t\t( 0 8 ) ( 1 9 ) ( 2 10 ) ( 3 11 ) ( 4 12 ) ( 5 13 ) ( 6 14 ) ( 7 15 )\n--------------------------------------------------------------------------------\nLevel:\t\t\t3\nSize:\t\t\t16 MB\nCache groups:\t\t( 0 8 1 9 2 10 3 11 4 12 5 13 6 14 7 15 )\n--------------------------------------------------------------------------------\n*******************************************************************************\nNUMA Topology\n********************************************************************************\nNUMA domains:\t\t1\n--------------------------------------------------------------------------------\nDomain:\t\t\t0\nProcessors:\t\t( 0 8 1 9 2 10 3 11 4 12 5 13 6 14 7 15 )\nDistances:\t\t10\nFree memory:\t\t4102.21 MB\nTotal memory:\t\t27760.4 MB\n--------------------------------------------------------------------------------<\/code><\/pre>\n\n\n\n
Following are the outputs from stream<\/p>\n\n\n\n
-------------------------------------------------------------\nSTREAM version $Revision: 5.10 $\n-------------------------------------------------------------\nThis system uses 8 bytes per array element.\n-------------------------------------------------------------\nArray size = 100000000 (elements), Offset = 0 (elements)\nMemory per array = 762.9 MiB (= 0.7 GiB).\nTotal memory required = 2288.8 MiB (= 2.2 GiB).\nEach kernel will be executed 100 times.\n The *best* time for each kernel (excluding the first iteration)\n will be used to compute the reported bandwidth.\n-------------------------------------------------------------\nNumber of Threads requested = 2\nNumber of Threads counted = 2\n-------------------------------------------------------------\nYour clock granularity\/precision appears to be 1 microseconds.\nEach test below will take on the order of 26015 microseconds.\n   (= 26015 clock ticks)\nIncrease the size of the arrays if this shows that\nyou are not getting at least 20 clock ticks per test.\n-------------------------------------------------------------\nWARNING -- The above is only a rough guideline.\nFor best results, please be sure you know the\nprecision of your system timer.\n-------------------------------------------------------------\nFunction    Best Rate MB\/s  Avg time     Min time     Max time\nCopy:           71476.8     0.023065     0.022385     0.025265\nScale:          70360.2     0.023345     0.022740     0.025791\nAdd:            73617.1     0.033653     0.032601     0.036008\nTriad:          73017.4     0.033967     0.032869     0.036821\n-------------------------------------------------------------\nSolution Validates: avg error less than 1.000000e-13 on all three arrays\n-------------------------------------------------------------\n<\/code><\/pre>\n\n\n\n
Following are the outputs from lmbench<\/p>\n\n\n\n
\"stride=1024\n0.00098 0.782\n0.00195 0.782\n0.00293 0.782\n0.00391 0.790\n0.00586 0.784\n0.00781 0.787\n0.01172 0.789\n0.01562 0.785\n0.02344 0.782\n0.03125 0.785\n0.04688 1.480\n0.06250 2.746\n0.09375 1.243\n0.12500 2.746\n0.18750 2.745\n0.25000 2.744\n0.37500 3.043\n0.50000 3.041\n0.75000 3.084\n1.00000 6.447\n1.50000 7.525\n2.00000 7.670\n3.00000 6.277\n4.00000 7.367\n6.00000 8.717\n8.00000 7.840\n12.00000 8.516\n16.00000 11.334\n24.00000 17.369\n32.00000 19.499\n48.00000 20.862\n64.00000 20.587\n96.00000 20.102\n128.00000 20.390\n192.00000 20.821\n256.00000 20.434\n384.00000 21.656\n512.00000 20.342\n768.00000 20.997\n1024.00000 20.326\n1536.00000 21.571\n2048.00000 19.647\n3072.00000 21.259\n4096.00000 23.743\n6144.00000 20.942\n8192.00000 21.334\n12288.00000 21.164\n16384.00000 20.588\n<\/code><\/pre>\n\n\n\n
Following are selected entries from output from lshw<\/p>\n\n\n\n
Motherboard: SER\nBIOS: American Megatrends: SER7PRO_P5C8V30 - 09\/15\/2023\nMemory: CT16G56C46S5.M8G1\nNVMe - WD_BLACK SN850X 4000GB<\/code><\/pre>\n\n\n\n
Following are several review and benchmark threads from phoronix.com<\/a> with 7840HS systems:<\/p>\n\n\n\n